ELF>@ @8 @kkPPP,<<bP|-== $$Std Ptd    QtdRtd,<<00GNUGNUjwj;-"ըB @,"  !L 4 "n@DD!04@$A!Ȓ&PA C(VP 19ufX> "0 (@`рIAT,aPDdP D @0 @ đ`( KH@"Dk$A!.CP: )PB@ @J"L0 A+h    "#$%&')*+,-./0124678:;<>@ABCEHIJKLMNPQRSTWXZ[\^`abcdefghijlmopqrstuvx{|}~UD9p3,_êA*S[DtkhM윙έ1;?1;9IJ%osW'gl Q{`~3^fH<'(#XMI~ (gxQw;!sD\~!sY+*`GYTuCH頡 w~ J%U&K%T9H nUX#L^QeFܽ'L^:> ݭ61DuUQGgB}61KfmmB}H/ gv.P '"K%] }tJ^CYd讋9F{Oqғ w\ mqs@e$; gRڃ1`Rvr a{vr!piI}WK%J%mI}Q_eA g H)g 6Ojm 6#GV{IVSmMV{Gg#kΑPe]=Mί1zڟ#1zؔΆ #__=K% 2Jŋ_{鵣£2J/={K`sO`ș˥}v-,_.,_;*[(՛ (#Aj -3 4pST  |od.weQ Ye"Dg >c$Nf $|J62(  9*Z O\qJ 7C x; /\o C)Io81jNMmUa_ RdB7x  C!pA y=un-,"( O\S4 Hy, #x& ^36  F" A.# np :   %  P) Y  pB75R@ 0$ P_= 8 Z) p! M P^ @ $ $ @#A ZJ   @ L')  V-=@ 00,30 ~(7)`L f06|"# C0A9@&3I&*W`00+% ( p9y>*0! )U0 Dp3 a'5E (P W" P); @(8E 5n +O` 0)  :Pu` u ` IU0+p[X X'/) o `Wm`pT5 ``! k P*@;  +`P+t` o`"Xp  h 0 P,, &~p `1/-0.m  Da  0 )m v !)  Q  *FP B! V @lP /: *3P  /  q7n/ '@CDI /   N  0$x ? 1PI U0" 1t p[J+ b@"q%4 &  03/2p!  ` _p  1$! 5 p   Y*- z+!  47X `  C2*++ +0 a $ ,P! `8 @iP 9!9[ p%B&1 p*a p) mp+   P=p0 *! I P P Lt0 i@y '0* `(  a`,0    /! @!   MI_@ ` ~@ ;   # p *PVj(7H&$ )9p3P$9tP(8O @@  j  H " 0A__gmon_start___ITM_deregisterTMCloneTable_ITM_registerTMCloneTable__cxa_finalize_Py_NoneStructPyDict_New_Py_FalseStructPyDict_SetItemPy_EnterRecursiveCallPy_LeaveRecursiveCall_Py_DeallocPyErr_OccurredPyObject_CallPyObject_GetAttrPyExc_SystemErrorPyErr_SetStringPyMethod_TypePyExc_TypeErrorPyErr_FormatPyDict_Size_PyDict_GetItem_KnownHash__stack_chk_failPyCapsule_IsValidPyCapsule_GetPointerPyErr_ClearPyNumber_AddPyNumber_InPlaceAddPyList_NewPyDict_TypePyTuple_NewPyExc_AttributeErrorPyErr_ExceptionMatchesPyObject_GetItemPyExc_ImportErrorPyObject_IsInstance_Py_TrueStructPyObject_IsTruePyLong_FromLongPyObject_SetItemPyFloat_FromDoublePyList_TypePyTuple_TypePyObject_SizePyFloat_TypePyFloat_AsDoublePySequence_ContainsPyLong_FromSsize_tPyObject_SetAttrPyObject_RichComparePyBool_TypePyLong_TypePyUnicode_TypePyUnicode_FormatPyNumber_RemainderPySlice_NewPyLong_AsLong_PyThreadState_UncheckedGetPyFunction_TypePyCFunction_TypePyType_IsSubtypePyNumber_InPlaceTrueDividePyBaseObject_TypePyNumber_SubtractPyExc_OverflowErrorPyObject_GetIterPyExc_ValueError_PyType_LookupPyEval_SaveThreadPyEval_RestoreThreadPyErr_SetObjectPyTuple_PackPyThreadState_GetPyNumber_LongPySequence_ListPyNumber_MultiplyPyList_AsTuplePyList_AppendPySequence_TuplePyDict_CopyPyObject_GC_UnTrackPyObject_GC_IsFinalizedPyObject_CallFinalizerFromDeallocPyObject_GetAttrStringPyDict_SetItemStringPyInterpreterState_GetIDPyModule_NewObjectPyModule_GetDictPyExc_RuntimeErrorPyOS_snprintfPy_GetVersionPyErr_WarnExPyFrame_TypePyBytes_FromStringAndSizePyUnicode_FromStringAndSizePyImport_AddModulePyObject_SetAttrStringPyUnicode_InternFromStringPyUnicode_DecodePyObject_HashPyLong_FromString__pyx_module_is_main_numpy__random__mtrandPyImport_GetModuleDictPyDict_GetItemString_Py_EllipsisObjectPyCode_NewPyType_ReadyPyCapsule_NewPyImport_ImportModulePyCapsule_TypePyExc_ExceptionPyType_ModifiedPyCMethod_New_PyDict_NewPresizedPyInit_mtrandPyModuleDef_InitmemcpyPyUnicode_New_PyUnicode_FastCopyCharacters_PyUnicode_ReadyPyObject_Formatrandom_standard_uniform_fillrandom_uniformPyFrame_NewPyEval_EvalFrameExlegacy_standard_exponentiallegacy_exponentialrandom_laplacerandom_gumbelrandom_logisticlegacy_gausslegacy_normallegacy_standard_cauchylegacy_waldrandom_triangularlegacy_standard_gammalegacy_gammalegacy_chisquarelegacy_flegacy_standard_tlegacy_weibulllegacy_paretolegacy_powerlegacy_lognormallegacy_betalegacy_logserieslegacy_vonmisesPyExc_NameErrorPyImport_ImportModuleLevelObjectPyDict_GetItemWithErrorPyExc_KeyErrormemcmpPyErr_NormalizeExceptionPyException_SetTracebackPyExc_StopIterationPyErr_GivenExceptionMatchesPyDict_Next_PyObject_GetDictPtrPyObject_NotPyTraceBack_HerePyUnicode_FromFormatPyUnicode_AsUTF8PyCode_NewEmptyPyMem_ReallocPyMem_MallocPyModule_GetNamePyCapsule_GetNamePyExc_DeprecationWarningPyErr_WarnFormatrandom_standard_uniform_frandom_standard_uniformrandom_standard_uniform_fill_frandom_positive_int64random_positive_int32random_positive_intrandom_uintrandom_loggamlegacy_random_hypergeometriclegacy_noncentral_chisquarelegacy_noncentral_flegacy_negative_binomiallegacy_random_poissonrandom_binomial_btperandom_binomial_inversionlegacy_random_binomialrandom_geometric_searchlegacy_random_zipfrandom_intervalrandom_bounded_uint64random_buffered_bounded_uint32random_buffered_bounded_uint16random_buffered_bounded_uint8random_buffered_bounded_boolrandom_bounded_uint64_fillrandom_bounded_uint32_fillrandom_bounded_uint16_fillrandom_bounded_uint8_fillmemsetrandom_bounded_bool_filllegacy_random_multinomialnpy_sinlnpy_coslnpy_tanlnpy_sinhlnpy_coshlnpy_tanhlnpy_fabslnpy_floorlnpy_ceillnpy_rintlnpy_trunclnpy_sqrtlnpy_log10lnpy_loglnpy_explnpy_expm1lnpy_asinlnpy_acoslnpy_atanlnpy_asinhlnpy_acoshlnpy_atanhlnpy_log1plnpy_exp2lnpy_log2lnpy_atan2lnpy_hypotlnpy_powlnpy_copysignlnpy_modflnpy_ldexplnpy_frexplnpy_cbrtlnpy_sinnpy_cosnpy_tannpy_sinhnpy_coshnpy_tanhnpy_fabsnpy_floornpy_ceilnpy_rintnpy_truncnpy_sqrtnpy_log10npy_lognpy_expnpy_expm1npy_asinnpy_acosnpy_atannpy_asinhnpy_acoshnpy_atanhnpy_log1plegacy_rayleighlegacy_random_geometricrandom_standard_exponentialrandom_standard_exponential_fillrandom_exponentialrandom_paretorandom_weibullrandom_powerrandom_rayleighrandom_standard_exponential_inv_fillrandom_standard_exponential_inv_fill_frandom_standard_normalrandom_standard_normal_fillrandom_standard_gammarandom_gammarandom_negative_binomialrandom_betarandom_chisquarerandom_frandom_normalrandom_lognormalrandom_standard_cauchyrandom_standard_trandom_noncentral_chisquarerandom_noncentral_frandom_waldrandom_vonmisesrandom_logseriesrandom_geometric_inversionnpy_exp2npy_log2npy_atan2npy_hypotnpy_pownpy_copysignnpy_modfnpy_ldexpnpy_frexpnpy_cbrtnpy_sinfnpy_cosfnpy_tanfnpy_sinhfnpy_coshfnpy_tanhfnpy_fabsfnpy_floorfnpy_ceilfnpy_rintfnpy_truncfnpy_sqrtfnpy_log10fnpy_logfnpy_expfnpy_expm1fnpy_asinfnpy_acosfnpy_atanfnpy_asinhfnpy_acoshfnpy_atanhfnpy_log1pfrandom_standard_exponential_frandom_standard_exponential_fill_frandom_standard_normal_frandom_standard_normal_fill_frandom_standard_gamma_frandom_gamma_fnpy_exp2fnpy_log2fnpy_atan2fnpy_hypotfnpy_powfnpy_copysignfnpy_modffnpy_ldexpfnpy_frexpfnpy_cbrtfnpy_heavisidefnpy_rad2degfnpy_deg2radfnpy_log2_1pfnpy_exp2_m1fnpy_logaddexpfnpy_logaddexp2fnpy_heavisidenpy_rad2degnpy_deg2radnpy_log2_1pnpy_exp2_m1npy_logaddexpnpy_logaddexp2npy_heavisidelnpy_rad2deglnpy_deg2radlnpy_log2_1plnpy_exp2_m1lnpy_logaddexplnpy_logaddexp2lnpy_gcdunpy_lcmunpy_gcdulnpy_lcmulnpy_gcdullnpy_lcmullnpy_gcdnpy_gcdlnpy_gcdllnpy_lcmnpy_lcmlnpy_lcmllnpy_lshiftuhhnpy_rshiftuhhnpy_lshifthhnpy_rshifthhnpy_lshiftuhnpy_rshiftuhnpy_lshifthnpy_rshifthnpy_lshiftunpy_rshiftunpy_lshiftnpy_rshiftnpy_lshiftulnpy_rshiftulnpy_lshiftlnpy_rshiftlnpy_lshiftullnpy_rshiftullnpy_lshiftllnpy_rshiftllnpy_spacingfnpy_spacingnpy_spacinglnpy_nextafterfnpy_nextafternpy_nextafterlnpy_get_floatstatus_barrierfetestexceptnpy_get_floatstatusnpy_clear_floatstatus_barrierfeclearexceptnpy_clear_floatstatusnpy_set_floatstatus_divbyzeroferaiseexceptnpy_set_floatstatus_overflownpy_set_floatstatus_underflownpy_set_floatstatus_invalidnpy_fmodlnpy_fmodnpy_fmodfnpy_divmodfnpy_remainderfnpy_floor_dividefnpy_divmodnpy_remaindernpy_floor_dividenpy_divmodlnpy_remainderlnpy_floor_dividelPyEval_EvalCodeExPyExc_IndexErrorPyNumber_IndexPyLong_AsSsize_tPyUnicode_Comparelibm.so.6libc.so.6GLIBC_2.14GLIBC_2.2.5GLIBC_2.4GLIBC_2.35GLIBC_2.27GLIBC_2.29@&ui 1ii =GR]ui 1<PQ<Q<T<U<U<U=U=UFFF FȢ F F G G GȢ G8 0GȢ @GȢ PGp XGȢ G G GȢ GȢ GȢ G G G H H H HȢ @H8 HHp PHȢ `Hp hHȢ H8 H0 HȢ H Hp HȢ H Hp HȢ H Hp HȢ I IȢ I (IȢ @I HIȢ `I` hIX pIȢ Ip IȢ IȢ Ip I IȢ Ip IȢ J J J JȢ @J HJ PJȢ `J@ hJp pJȢ J@ JȢ J Jp JȢ JȢ J J JȢ K K KȢ K8 (K 0KȢ 8Kx @Kp PK XKȢ K Kة KЩ KȢ K KȢ K KȢ K KȢ L0 (L 0LȢ @L0 HL PLȢ S aS  S @T aHT Q`T ahT ST aT tT (UT {T `P T -UT pT K T aT T D U TU UU  > U T(U @W8U `= @U THU ZXU 7 `U ThU ^xU . U TU `aU * U |UU  U @% U 7UU U  U aU U ` V VUV V  V V(V 8V  @V aHV FXV  `V  WhV IxV  V WV 0MV ` V PVV V  V YVV V  V `VV V  W iVW W  W zV(W  8W  @W oVHW XW  `W VhW xW  W |VW W  W VW W { W VW W `j W VW W ] X VX X M X V(X 8X ? @X VHX XX  2 `X VhX xX `$ X VX 0X ` X VX X   X VX X X VX @X Y VY  Y @ Y W(Y @ 8Y @Y [HY XY `Y [hY bxY Y [Y pfY Y [Y iY  Y [Y lY Y [Y sY Z [Z oZ  Z [(Z Й8Z n@Z aHZ LXZ a`Z (^hZ pxZ UZ aZ `mZ  QZ aZ ,Z `LZ xZ  [  ([ zH[  P[ pzp[  x[ @z[  [ 0z[ ȫ [  z[ H [ y\  \ y8\ p @\ y`\ @ h\ y\  \ y\  \ y\  \ x] P ] x(] h 0] xP] P X] xx]  ] Px] @ ]  x]  ] w]  ] w^ ` ^ w@^ p H^ pwh^  p^ @w^  ^ w^  ^ v^  ^ v_  _ v0_  8_ @vX_  `_ v_ Э _ u_  _ u_ @ _ @u_ 0 ` u ` P (` tH`  P` tp`  x` pt`  ` @t` 0 `  t` Ю ` sa Я a s8a  @a s`a  ha sa  a `sa @ a  sa  a rb P b r(b  0b rPb  Xb `rxb 0 b  rb  b qb  b qb  b `qc  c @q@c ` Hc qhc  pc pc  c pc p c pc  c `pd  d @p0d  8d  pXd  `d pd  d pd H d nd  d nd Ч e `n e  (e nHe  Pe mpe ` xe me ث e me  e me  e mf  f m8f  @f m`f  hf mf  f mf ج f `mf  f  mg  g l(g  0g lPg  Xg lxg  g |lg  g xlg ` g plg  g `lh  h Sl@h P Hh Hlhh  ph Blh @ h ;lh  h 6lh H h 0li  i (l0i 0 8i lXi  `i li  i li  i li  i li 8 j l j  (j lHj  Pj kpj  xj kj X j  ]j ` j ]j p j \k  k \8k  @k \`k  hk Zk @ k @Zk  k 0Zk  k 0Zl  l (Z(l x 0l ZPl ` Xl Yxl  l Yl  l Yl H l Pl ص l Pm ر m P@m  Hm Chm  pm Bm ` m Bm  m Bm X m Bn  n B0n  8n BXn  `n Bn p n `Bn X n  Bn ( n Bn p o B o h (o BHo  Po Bpo  xo Bo  o Bo  o Bo  o Ap б p A8p ȭ @p 5`p  hp 5p  p 5p 8 p 5p H p 5q  q 5(q X 0q x5Pq  Xq o5xq ئ q i5q ` q `5q  q P5q ȱ q P5r  r K5@r  Hr K5hr h pr `'r  r X'r  r P'r ȣ r D's x s >'0s  8s >'Xs x `s `s  s Zs P s Ps  s Ps  t  t  (t Ht  Pt pt  xt t 0 t `t  t Xt 8 t Pu  u P8u  @u ``u  hu Pu  u Gu  u 8u  u 8v  v (v  0v Pv  Xv xv  v v  v v С v v  v w @ w @w  Hw hw 0 pw w  w w ` w w  w x 8 x 0x  8x xXx  `x qx  x hx X x ax P x ax  y ] y ( (y VHy  Py Tpy  xy Py  y Py  y Hy  y Hz Ȯ z 8z  @z `z  hz z  z xz  z pz  z `{  { S({ h 0{ HP{  X{ @x{ @ { @{ أ { ;{ P { 0{  {  |  |  @|  H| h| ش p| |  | |  |  |  | } x } 0}  8} @X} P `} 5}  } 1}  } (}  } } 8 ~  ~  (~ H~ h P~ p~ x x~ ~  ~ ~  ~ ~ h ~    p8 ` @ l` H h l   ` p  ` ح  @    ( h 0  P  X x 0 y  yȀ  Ѐ p ة g  g@  H bh x p P ` P H      0  8 X  ` Њ  Ŋ x ŊЂ ` ؂  X  8 ( ~H H P ~p P x ~ X q 0 ȃ eq H eq   b8 @ @ a` Щ h a  a x a؄ 8 a 0  a( ` 0 haP  X Pax  Ha ` `ȅ  Ѕ p`  j`  j`@  H P`h  p F` @ F`  P  XP 8  XP0  8 KX h ` K  K 0 KЇ 8 ؇ A X A  ( AH س P ~Ap ( x ~A خ 3 Ȭ Ȉ 3 Ь P3   038 ؠ @ 3`  h 3  3 p 3؉  2 ؟  2(  0 2P  X 2x د . 8 h.Ȋ  Њ h.   #  #@  H #h ȯ p      p    0  8 `X  ` H  H H Ќ p ،     ( H  P p  x y H p  ȍ p   H  `8  @ X` 0 h P ȟ @ Ы 0؎  '   '(  0 "P  X x   p ȏ h Џ     @ ذ H h X p  x ` а `    @  0  8 X 8 `  Ȱ   Б  ؑ ` 0 Z ( ( ZH Ȣ P Up  x P  K ȷ Ȓ @  0   08 ( @  `  h    8 ؓ в    (  0 P ا X x     Ȕ  Д     @ H H h  p  ( p  @ У 8 (  00  8  X и `      Ж h ؖ      ( H  P p p x  (   ȗ    h   8  @ ` ( h      ؘ     ( x 0 P  X x   П ș ( Й  `   @  H h H p    x  0 q   q0 8 8 hX 0 ` ` x `  ~Л @ ؛ `~ p @~  ( 4~H  P /~p p x /~ ( s 8j0 PKX m q Xj     @T S ( c8   a^ @H R F^Ȟ 0E؞ `S  a  ( Z P  X Ȣ 0  Ȣ >>>>??`?8?C ?E(?F0?G8?l@?pH?rP?vX?`?h?p?x????????????????@ @(@0@8@@@H@P@ X@ `@ h@ p@ x@@@@@@@@@@@@@@@ @!@"A#A$A%A& A'(A(0A)8A*@A+HA,PA-XA.`A/hA0pA1xA2A3A4A5A6A7A9A:A;A<A=A>A?A@AAABADBHBIBJBK BL(BM0BN8BO@BPHBQPBRXBS`BThBUpBVxBWBXBYBZB[B\B]B^B_B`BaBbBcBdBeBfBgChCiCjCk Cm(Cn0Co8Cq@CsHCtPCuXCw`CxhCypCzxC{C|C}C~CCCCCCCCCCCCCDDDD D(D0D8D@DHDPDXD`DhDpDxDDDDDDDDDDDDDDDDDEEEE E(E0E8E@EHEPEXE`EhEpExEEEEEEEEEEEEEEEEEFFFF F(F0F8F@FHFPFXF`FhFpFxFFHH9_HtH5_%_hhhhhhhhqhah Qh Ah 1h !h hhhhhhhhhhqhahQhAh1h!hhhh h!h"h#h$h%h&h'qh(ah)Qh*Ah+1h,!h-h.h/h0h1h2h3h4h5h6h7qh8ah9Qh:Ah;1h<!h=h>h?h@hAhBhChDhEhFhGqhHahIQhJAhK1hL!hMhNhOhPhQhRhShThUhVhWqhXahYQhZAh[1h\!h]h^h_h`hahbhchdhehfhgqhhahiQhjAhk1hl!hmhnhohphqhrhshthuhvhwqhxahyQhzAh{1h|!h}h~hhhhhhhhhqhahQhAh1h!hhhhhhhhhhhqhahQhAh1h!hhhhhhhhhhhqhahQhAh1h!hhhhhhhhhhhqhahQhAh1h!hhhhhhhhhhhqhahQhAh1h!h%RD%RD%RD%RD%RD%RD%RD%RD%RD%RD%RD%RD%RD%RD%RD%}RD%uRD%mRD%eRD%]RD%URD%MRD%ERD%=RD%5RD%-RD%%RD%RD%RD% RD%RD%QD%QD%QD%QD%QD%QD%QD%QD%QD%QD%QD%QD%QD%QD%QD%QD%}QD%uQD%mQD%eQD%]QD%UQD%MQD%EQD%=QD%5QD%-QD%%QD%QD%QD% QD%QD%PD%PD%PD%PD%PD%PD%PD%PD%PD%PD%PD%PD%PD%PD%PD%PD%}PD%uPD%mPD%ePD%]PD%UPD%MPD%EPD%=PD%5PD%-PD%%PD%PD%PD% PD%PD%OD%OD%OD%OD%OD%OD%OD%OD%OD%OD%OD%OD%OD%OD%OD%OD%}OD%uOD%mOD%eOD%]OD%UOD%MOD%EOD%=OD%5OD%-OD%%OD%OD%OD% OD%OD%ND%ND%ND%ND%ND%ND%ND%ND%ND%ND%ND%ND%ND%ND%ND%ND%}ND%uND%mND%eND%]ND%UND%MND%END%=ND%5ND%-ND%%ND%ND%ND% ND%ND%MD%MD%MD%MD%MD%MD%MD%MD%MD%MD%MD%MD%MD%MD%MD%MD%}MD%uMD%mMD%eMD%]MD%UMD%MMD%EMD%=MD%5MD%-MD%%MD%MD%MD% MD%MD%LD%LD%LD%LD%LD%LD%LD%LD%LD%LD%LD%LD%LD%LD%LD%LDAVIAUIHATUSDHt/H;;EHuE1tHLL@AH蕢HdEAH8tE1[D]A\A]A^AVAUATUHQHx[H,Y HuHY Ht#H9tHDH5gH8E1L% Mt I$H5wcHIHtHYLIΡMtL1IHAH 3cHHH1cxkAH "cLHHcxHAH cLHHcpx%E1H cLHHbPxMLZL]A\A]A^AWAVAUATUHSHH dH%(H$1Ht'H9[QHBH5fH8}=QLl$(A 1HbLH\$,5HhbHH1D$,8D$(u D$.8D$*t?Ld$01IHbfLLdL1LKHB1H@ HH H HKL%>~1L H HK1LHB HKH-R HHEH HKHH=qaH HKHH=[aH H|KHHޭ H5DaA\H=ڭ AHe] `L;MtmC C!H{t.{"t.HCHSHpHt 1( HCHpIHH8HNBHNH(WHD HNWH HNn|H HNY|H% HgN1HV HPNH H6N11H= `H HNHH HMH?8HHIL%:bHLgHH= 蒻H{ HA\AH= tAH= tH UDDH=aH=ګ Ht2Hʫ  6HuH?H5aH8H= tMH H5 H= A\ASHZ LHGA\A(H= 蒺HS HH= vHϢ HH=g ZH HH=C >HH=& )H HH= H H{H=ު HR H_H= չHJH= Hٵ H.H=y 褹H5 HH5U 1H HuA\AH5 1HB HtH5 1H HtHg H5X 1H= HtH51 1mH HhH5 1JHà HEH5۪ 1'H H"H0 H5 1H^ HH H5 1H, HL%<1LLLH HH H5) 1}H HxH H5 1SH HNH5̩ 10H H+H5 1 Hf HH5v 1H3 HH5K 1H HH5 1H HH5 1Hz H|H5ʨ 1^H? HYH5 1;H H6H5t 1Hٛ HH5I 1H HH5 1H[ HH5 1H( HH5 1H HH5 1iH HdH5z 1FHϡ HAH5O 1#H HH5$ 1Hy HH5 1H. HH5Φ 1H HH5 1Hh HH5x 1tH= HoLLLH HQH5/ 13HĢ H.H H5 1 Hj HH5ڥ 1H? HH5 L1H HH5 1H HHn H5W 1sH HnHD8H5 1IHr HDH5 1&Hǯ H!H5פ 1H HH5 1HH H5y 1HH L Ʈ VA111AQh65$ 5. RRPRRHPHuH H5 1PHRHP L Y QA111AQh=5 5 RRPRRHPHH H H HHHHǤ HdH DH=ݭ 11L%B HHDH5 LHH]<H5P H= H@DL%UH+ LHH DAxLHHULDH;HyLHHCA LHHV DH;H7LHHCA LHH(UCHa;HL%UL6HH@CA`LHHTCHޗ H;AH LHHTRCH:A0LHHT,CH H:AXLHHTBH H:ALHHaTBHk:ALHHCTBHE:ALHH,TBH5 H:ALHHSYBH9ALHHS3BH9ALHHS BH9ALHHSAH Hy9ALHHpSAHS9ALHHZSAH-9ALHH=SnAH9ALHH!SHAH8HL-SnLHH@A`LHHgH@H8A@LHHR@Hr8HH5 IHS81HHHu HuH1H55VH8}L赏H8ALHHXRQ@H7HL-QRwLHH?L%FRHը HLH5@RA7LHџ H5RH@q7H RH HH5R@L7HH=~U1HH?L%UH HLH5Q~A6LH[ H5QH]A6LH* H5QH LHH,LH5 H=l H&HHM H H9Xu#H-ќ HtHE7H= ʡHH=~ H H5 +HH0H5 H`IH[&HlH5] H=Ɗ Lv@,LFH H( H9Xu#L% MtI$7H= $IH=؜ H H5 腡IM0H5› L躠HH+LƂH5 H= H%H蠂H Hj H9Xu#H-U HtHE7H=C ~HH=2 H+ H5, ߠHHn/H5 HIH4%H H5 H=z L*+LH[ H H9Xu#L% MtI$7H= ؟IH= Hm H5n 9IM.H5F LnHH*LzH5# H=Ԉ H|$HTH H H9Xu#H-ٙ HtHE7H= 2HH= H H5 蓟HHD.H5 HȞIH $HԀH5e H=. L)L讀H H0 H9Xu#L% MtI$7H=Q 茞IH=@ H H5 IM-H5 L"HH})L.H5׆ H= H8U#HHi Hz H9Xu#H-e HtHE7H= HH= H; H5< GHH-H5 H|IH"HH5 H= L(LbHÆ H H9Xu#L% MtI$7H= @IH= H} H5~ 衝IM,H5V L֜HHS(L~H53 H=< H."H~H H H9Xu#H- HtHE7H=_ 蚜HH=N H H5 HH+H5 H0IH!H<~H5u H= LF'L~Hw H@ H9Xu#L%+ MtI$7H= IH= H H5 UIM[+H5ڕ L芛HH)'L}H5 H= H!Hp}Hф H H9Xu#H-m HtHE7H= NHH= HC H5D 诛HH*H5 HIH H|H5 H=J Ln&L|H+ HĔ H9Xu#L% MtI$7H=m 訚IH=\ H H5 IM1*H5^ L>HH%LJ|H5; H= HTH$|H H H9Xu#H- HtHE7H=Ǖ HH= HǓ H5ȓ cHH)H5 H蘙IHqH{H5} H= LD%L~{H߂ HH H9Xu#L%3 MtI$7H=! \IH= H H5 轙IM)H5 LHH$LzH5 H=X HHzH9 H H9Xu#H-u HtHE7H={ 趘HH=j HK H5L HHr(H5$ HLIHJHXzH5 H= Lb$L2zH H̑ H9Xu#L% MtI$7H=Փ IH=ē H H5 qIM'H5f L覗HH#LyH5C H= HHyH H H9Xu#H- HtHE7H=/ jHH= Hϐ H5А ˗HHH'H5 HIH#H yH5 H=f L"LxHG HP H9Xu#L%; MtI$7H= ĖIH=x H H5 %IM&H5 LZHH"LfxH5Ǐ H= HpkH@xH H H9Xu#H-} HtHE7H= HH=ґ HS H5T HH&H5, H贕IHHwH5 H= L!LwH~ HԎ H9Xu#L% MtI$7H== xIH=, H H5 ٕIM%H5n LHHW!LwH5K H=t~ H$DHvHU~ H H9Xu#H- HtHE7H= ҔHH= H׍ H5؍ 3HH$H5`y HhIHHtvH5=y H=} L~ LNvH} H` H9Xu#L%K MtI$7H= ,IH= H! H5" 荔IM_$H5 L“HH- LuH5׌ H=(} HHuH } H H9Xu#H- HtHE7H=K 膓HH=: Hc H5d HH#H5< HIHH(uH5 H=| L2rLuHc| H H9Xu#L%ϋ MtI$7H= IH= H H5 AIM5#H5w LvHHLtH5sw H={ HH\tH{ H. H9Xu#H- HtHE7H= :HH= H H5 蛒HH"H5Px HБIHHsH5-x H=6{ LHLsH{ H x H9Hu#L%c MtI$7H=Y 蔑IH=H H9 H5: IM "H5 L*HHL6sH5 H=z H@HsHqz H H9Xu#H- HtHE7H= HH= H{ H5| OHHv!H5T H脐IH`HrH51 H=y LLjrHy H H9Hu#L% MtI$7H= HIH= H H5 詐IM H5 LޏHHLqH5s H=Dy HHqH%y H> H9Xu#H-) HtHE7H=g 袏HH=V H H5 HHL H58w H8IH9HDqH5w H=x LNLqHx H H9Hu#L%s MtI$7H= IH= HI H5J ]IMH5" L蒎HHLpH5 H=w HHxpHw Hʆ H9Xu#H- HtHE7H= VHH= H H5 跎HH"H5Dw HIHHoH5!w H=Rw LLoH3w H  H9Hu#L% MtI$7H=u 谍IH=d HՅ H5օ IMH5v LFHH[LRoH5u H=v H\ZH,oHv H^ H9Xu#H-I HtHE7H=ψ HH= H H5 kHHH5 H蠌IHHnH5 H=v LLnHu H H9Hu#L% MtI$7H=) dIH= Hi H5j ŌIMcH5B LHH1LnH5 H=`u H3HmHAu H H9Xu#H-Ճ HtHE7H= 辋HH=r H H5 HHH5 HTIHH`mH5a H=t LjvL:mHt H , H9Hu#L% MtI$7H=݆ IH=̆ H H5 yIM9H5Ƃ L變HHLlH5 H=t H HlHs Hn H9Xu#H-Y HtHE7H=7 rHH=& H/ H50 ӊHHH5(w HIHHlH5w H=ns LFLkHOs H H9Hu#L% MtI$7H= ̉IH= Hy H5z -IMH5R LbHHLnkH5/ H=r HxHHkHr H H9Xu#H- HtHE7H= &HH=ڄ H H5 臉HHzH5 H輈IHvHjH5q H="r LLjHr H < H9Hu#L%' MtI$7H=E 耈IH=4 H H5 IMH5 LHHL"jH5 H=|q H,HiH]q H~ H9Xu#H-i HtHE7H= ڇHH= H? H5@ ;HHPH5 HpIHOH|iH5~ H=p LLViHp H ~ H9Hu#L%~ MtI$7H= 4IH= H~ H5~ 蕇IMH5Z~ LʆHHtLhH57~ H=0p HHhHp H~ H9Xu#H-} HtHE7H=S 莆HH=B H} H5} HH&H5} H$IH(H0hH5y} H=o L:L hHko H D} H9Hu#L%/} MtI$7H= IH= H} H5} IIMH5| L~HHDLgH5| H=n Hp HdgH| 11H=2i 轺HH6H5b| H=n HK8 HgHD| 11H=h tHHH5| H=Rn H Hf38HHH@H{ HHH{ HHPH{ HHPH{ HHPHr{ HHP H\{ HHP(HF{ HHP0H0{ HHP8H{ HHP@H{ HHPHHz HHPPHz HHPXHz HHP`Hz HHPhHz HHPpHz HHPxHjz HHHQz HHH8z HHHz HHHz HHHy HHHy HHHy HHHy HHHy HHHpy HHHWy HHH>y HHH%y HHH y HHHx HHHx HHHx HHHx HHHx HHHvx HH H]x HH(HDx HH0H+x HH8Hx HH@Hw HHHHw HHPHw HHXHw HH`Hw HHhH|w HHpHcw HHxHJw H5+w H= k HHH#w HHH w HHH落 H`c,FHHrHv H5v HTt Hv H5~v H6a H_v H5Pv HN H1v H5"v H; Hv H5u Hܼ( Hu H5u H込 Hu H5u H蠼 Hyu H5ju H肼HKu H5A9a6A:a.A;a&AaA?aA@aHWMLWzDDH xAZ]H=FAiPHWHWHWWA,]A2A;]A!AJ]AA]AA]AA]AA]AA]AA]A A]A A]A A^A wA^A fA"^AUA.^ADA@^A3AL^A"A^^AAj^AA|^AA^AA^AA^AA^AA^AA^AA^AxA^AgA_AVA_AEA_A4A0_A #A<_A!AN_A"AZ_A#Al_A$Ax_A%A_A&A_A'A_A(A_A)A_A*|A_A+nA_A,`A_A-RA`A.DA`A/6A `A0(A,`A1A>`A2 AJ`A3MMA\AöA\A鲶A\A顶A\A鐶A\AA\AnA\A]A\ALA ]A;A]A*A$]AA9]AAE]AAc]AA{]AյA]AĵA]A鳵A]A颵A]A鑵A]A逵A]A oA]A ^A]A MA]A A_A+-A_A,A_A- A_A.A `A/A`A0زA*`A1DzA9`A2鶲AH`A3饲AW`A6锲Ac`A=郲Ao`ADrAagA\AVA\AEA\A4A\A#A\AH$dH+%(t谢H[]A\A]A^A_H=CQ 鎠AVIHAUIATUDSHHdH%(H$1IHH@u H4LLH5-H81nML$ I9s#HILLH5-H81AuHI9vCHPMLAQH .IH11H1ĠHy LE1RNH$dH+%(t芡HL[]A\A]A^AWIAVIH5l,AUIATUSHAPHŃHLH赥IHu(LգLH5-HHH81רgLHu9LNLH蓣IMLHHH5-H81菨LL蒨HHt HRM1 HFMZ[]A\A]A^A_AWIAVIH5|+AUIATUSHAPHŃHLHŤIHu(LLH5[-HHH81gLHu9L^LH裢IMLHHH5A-H81蟧LL袧HHt HbL1 HVLZ[]A\A]A^A_H=N HN H9tH6Ht H=N H5N H)HH?HHHtHHtfD=N u+UH=Ht H=nٛd]N ]wHGPHGXHÐAVAUATUHHGH5Q HHHHhIHwHpH5Q H!HEL5Q LMH=BLLHAIfMtYHmt:ImtHL]A\A]A^@LHL]A\A]A^fH؜ImusHAHmuH襜ImDE1H H=luHL]A\A]A^DALLHrIHtHm%Zf.A{kHrHmA}eHXfLIHlH5H8AVAUATIUHHGH5O HH#HHHH9EuxLuMtoLmIIEHmtKLLLI.It|IEHIEMt{HtVI,$t?H0HH]A\A]A^HfDLHIIDLfDLКfDLwHtNH H=sH1]A\A]A^f軣HLt$ Tt$ ff.AVAUATUSHHL-6HndH%(HD$1L,$H HHHH\H MHHHHN H?L %HLHL@HHUH'H53H81IXZH vH=E1grHD$dH+%(LHL[]A\A]A^HVHH=RHs MHEHL IHEMt]HHEuH fLIHHHFHH$HGH$x@HHEt&H 8H=q@H舘1HL IH5Ly f.H舖IHoH5K LHVIHtH$IFEfAVAUATIUHHndH%(HD$HH$HHHLnID$H5K LHH2HHIHH5J LH!9HEL5K LMH=BLLHAIfMHmI,$Lf.HHZH KHHHHLH?L "HLHL@HHUH3H51H81GXZH H=ME1eoHD$dH+%(HL]A\A]A^fIH8!AHmuHI,$DE1H _H=nxDIHHHFHH$HjL,$@LLHIHyfDA[DAeD軞HLXC1HL IH5Lp|˙HuHH5H8PkH0IH'H5UH LHVHuH$IF薕HmAH蝔AWAVAUATUHSHH(L5LndH%(HD$1H$HD$Lt$H~I\IHV(HE LMHHHsHH=ZAHEHAVj58G 5*G j5G Pj5 G #G IHEHPMHHEHD$dH+%(H(L[]A\A]A^A_HH5F LIHVI膘H$HLmDIH HAHMEIHHHH5AUL H81貛XZH UH=E1k*LHȒII[qM IHFHH$ΐIH5E LHV蘗HD$H3IML $HD$HT$,f.IHF(HHD$HF HD$HFH$TIM~1HL MH5&LnyHHEt&H 0@H=j@H訑fDH A`fDH5D LHV蕖H`HD$IMfDHF HHD$HFH$IHH H H5jL AHH81讙Y^͑H ZAff.AVAUATUSHH L C L-dH%(HD$1HnL $Ll$HQH?H%HHHIH HHIHH"I?UIH5H8L A1ΘXZH qH=$E1hHD$dH+%(H L[]A\A]A^MHHC HHsHHB H=KAHEHAUjRPjRLPj5B B IHEHPM|HHEbHVULV LNvMIHHteHHTIHL $LT$%fHHEH <H=gHF HHD$HFH$H~1HL 6IH5Ly_fDHFHH$蠌IMTH5A LHVaHtHD$IF느H(AH59A LHV-HtH$IَfAVAUATUHSHHLfdH%(HD$1H$H%MIMH H HHLHH?L aMLHL@HHBATHH5H81XZH "H=~E1fHD$dH+%([HL[]A\A]A^fHVHHuHIE1LH=HI$LSjSSjSSjS@ II$HPMt^HI$uL褌yHIMIHFHH$誊HGH$`@HI$t&H FCH=!e@L 1HL MH5Ly fH(IHwH5M? LHVHtH$IFM蒌fAWAVAUATUSHHL-HndH%(HD$1L,$HHHulHnL9H5@ HFHEHHHHHJH/PHEHHEHHbH SHHHHTH?L HLHL@HHUH\H59H81OXQ ZH H==AjcHD$dH+%(HD[]A\A]A^A_@HA? H A H9HL%A MI$HI9D$LHHEI,$BHEH{HHEH/8HEHkHH5f> HHvIMxL-VLLfLLCIHAoEHC AoMK0IE HC@HC HCHHCHHH(tHEH5= HHHIMHH/GLAHmI,$ZL̈M軈fDL計蛈fDL-= L5= IEHHNH=bS1LLI舍MsLAA-ImXH zDDAH=`HmfDHM`f.kH< H ? H9HKL-> MsIEH/I9EuLuIMImIFH5 < LHHIMHI9EM}MMeII$ImlHLLI/HHXI,$HmIL軆fD諆fDIHHHFHH$躄HH,$HLMHYf.苏I H H=A^HmE1 f.L1HL IH5]LELC L訅5KHAA?aAARO賎IJAA]'HH5H8z H zH=A]fH!LHIH[H59 LHVىHH$IF)L蘄H=A9 H; H5; PIMG ;裍H(Mt$M)Ml$IIEI,$+LLMI.HLH AImMH 9DAH=~\B@I,$}H  H=M\A DH=18 OI@H=8 H: H5: nOIM H=7 NIM}M~MeII$ImLLMI/IVLЂILL fDIHI,$ ALrH DH=J[MI.AL9@E1 AL1L3IHAA)L ALсwAA)HH;H5tH8wrME1 H  H='ZZDAWAVAUIATUSHHGH55 HH0HH:HH9]YL}MLLuIIHm%LLI/IM|I.RH=+5 H5,5 HGHH HHH=4 L#IHH9]L}MLuIIHmLLLI/HImHI.HLuII$MHI$H+HL[]A\A]A^A_HHIIDLHI2ImHaL|TLhLXCHHH8VL(<L>LKHHADE1H 1H=WI$AHI$uL~HtHmuH~fDˇHI$AHI$uf.I.A^L8~QHI$tFH+tA3DH~fDI$LAHI$AIL}fDAUH5#2 ATUHGHHHIMID$H51 LHH,II$HI$M"HH}H51 HGHHbHHHEH5y1 HHHFIHEHM@HEHH=41 LL|HH0I,$H51 H&|IHHmLL{HHt|I,$ImHEIHPHUHEHL]A\A]ÐHH|XL8|L(|^H|t,I,$uL{LE1H IH=TIEHhH{L]A\A]L{-L{ۄIMH H=RTL]A\A]f蛄IHH E1H=4TfA!DLE1H VH=SIEHu@#HIA#HmuHz뗐&fDA)LzDff.AWAVAUATUSHHGH5. HHIM谀HHH H5N. H辁vIUL%+. HHH=xHLLIMImHmwIHH- IT$LHHH=- hGHI$HHI$HEL%_- HHLHIMI}HmH5- H/LI9FuHIM|HHCL`|IHIELhHX IMt$(IHHImtkHL[]A\A]A^A_HAA/I$fDDIH H=w"QI.AMtImuLx@HxVLwHm@|H"AImQHmH DE1H=PAI,$H CH=HPI.AL8wL(wHwLwLIAAHfDLvHLL~IHImfDH H=E1xOI5LE1A_vHmH DDH=+OMME1ImmLt$ vt$ XE1E1AHuH >DDH=NMMf.H9H5rH8vDE1~IWHH8ywMAA2DH AH=pNI.@vI)H+H XHH=&M@H 2*H=MI.0Lt#@HLMH5NAA2H81&}LAMACmfAWAVAUIATUSHHDHndH%(HD$1H$HHHPLvImH( H 2+ HEH9HL%+ M;I$LHuAHEAHHEI,$EI}H5%( HGHHIMHI9D$IM|$M;Il$IHEI,$LLHI/I'HEHMnHEHII./IELHHZH(@HH~@HHH HHHH H?L oHLHL@HHPUH@H5H81{XZH H=(K1HT$dH+%(0H[]A\A]A^A_I LLLIf.LqLq2HqLqHqLqa{qfDIH HHFHH$oH\L4$<H=% Hb( H5c( 6=IMDAAHmuHpH BDDH=DI1HHEI,$AAuLpDH=A% $<Ii@H-q$ L%r$ HEHHtOH=NouT1HLHxuHHAA 軼Hm(H1LxHHuAA1HL IH5OLG~eDAAxI'A)AyAA5wHmIHH5# LHVQtHCH$IFsHHH5H8GpoHn.AWAVAUATUSHH8L-ܾHndH%(HD$(1Ll$ HbHH7HnH{HEH5s" HGHHIMH5b" HLI9D$=IMH5" LH>ADžI.E+L9H;-ݽbH;-UHsH" H $ H9HL5$ M-IIFH5f! LHHIIHMIHRIGH5! LM_H=Ht$&l^1Ht$LAIIrMI/I.<HռHHmH-Hc{PpIHBH5 HLjPI.FCXloIHH5A HL^jI.L9H;-?H;-MHr/H5( HILI9D$#;IM"H5 HLI9D$a;IMH5v HLI9G5;IMI//H5H HLI9D$K:IM"HI9CL\$LH5 9:L\$HHI+H5 HLLI9D$:IM H5 HLI9D$d:IMLT$nLT$HHX(I$LILpLh HSLx0LP8fHHHmHaj@HHH HHHHH?L HLHL@HHUHH5H81rXZH RH=}E1BHD$(dH+%(-H8L[]A\A]A^A_LqLi'Lhi`[ifDLHiIHHHFHHD$ YgHHl$ {fDLhI$LHBI$\@LhHhH-@Lh H H= sAMf.I$LE1HPDqI!@iI!E1E1I.dLT$Mt$L\$gt$T$E1H\$fDMtImtDHtH+taMtI*t~H=OH @I$LE1HPLT$t$LT${gT$t$LT$@H߉T$t$LT$SgT$t$LT$xL׉T$t$0gT$t$H H=@1HL$ ILL H5epyfD1LnIHf"E11E1I/LT$t$LT$fLT$t$T$D&fLXf&E1E1#HpdIHH5U LHV1kHHD$ IFf.H= H H5 1IM9"%'fLeqE1۾"ffDnIH=I 0I'E1E1#(f!{iHBHH5H8AŃf HmkH{DkPLH5 HGHHm Ѕ{L |IMI,$Mt I.HL[]A\A]A^A_fH59 L`YH- L% HEHHAH=]F1HLH dH HZHmcufHt$ $^$t$ H H=DE1d7$ID$H(HEPfDH5  L` I$E1&fDH- L% HEHHH=\1HLHcHHgHmg( DMl$IE#fL])Lx](Hh]Il$HEKAEDID$L(IEfDHm kE1E1E1AMImtJMtI(t/H QDH=5M|ME1fL\fDLL$\L$fDL\BH\L$s\$ fID$L@I]fDL@\yH0\1^IHDH jh&E1H=4L[A(hE1E1HmE1E1HL$[ML$1LHcIHMrAMD$ IafID$L@IfDA[lHmAA [HL\ImHLZ_IHH 3gE1H=3i@H1LbHHH cE1H=Dg3'fH- L% HEHH#H=BY(1HLHh_HH跦Hm7ix]DtH HH=E12n1LLaHHImuLYH sMH=Nq2@H ҾkIE1H=%H2LHY1 \IHHLZI/IL YMD$(I|fA_lE1E1AKkf.b'raHLX^E1E1AWlf.ID$H;aH;H@hH<H@H/LIMUH5 LHUI.ID$H; H;@H@hH`H@HSLIMH5_ LH|U_I.HEIIjf.AakE1 D$[$HImuL!WH {rE1H=κ/@sf+`I{[MrAHH=H5vH8WfVYIHCHLWI.IhLH$VVL$SDZHMsA`AYlZfDZHBH{H5H8,W'XIHHL/WI.I#LH$UL$DA]lKbH1Lt]HHyjImmE1=H1L>]HHH li3E1H=-ID$LpILTID$Lp ILT5Mt$0IE1AKkMt$8I+YIH]HH51H8UiH 3H= ,-E1AYlIAzLnE1DE1A]lVIH HLWUImILSInxUXIHH"H5[H8TcH ظH=3V,'VIHHLTImIL*S~IALoE1EIo HH5H84TgH 9H=+wE1InxE1IoE1E1E1kAMff.@H/tD[Rff.AWAVAUATUHSHXHDLnH|$dH%(HD$H1H H\$(HD$ H\$0H$HD$8H'IwII_ H\$Le LmIEI$I9;H4$H=X ;HHWH@H54 HHH IML;5tL;5BI9LWADžAE1E1I.LDL$ML$PL$DL$Mt I(H ;DAH=)Hm1MgI.]LPImULPI,$MLpP9IvHF0H$HE(HD$DI.jH# HJEyH9  L5  MIIFH5 LHHIIHMIHIGH5[ LMCH=LHt$N 1Ht$LAITM,I/I.HEH5 HHHHHHH9G[LGMNLIIH/ LLLD$舟LD$II(IHMIHZ Hm7 H LH H9B L5} M\ IIFH5  LHHR IIHM IH1LHNIH5 I/3L;5,L;5u I9DI.EH H  H9H;L= MIIGH5 LHHiIIHM<IHx LHNIHtI.r L;=cL;=1u I9a DI/c E H  H  H9HML5 MIIFH5cLHHsIIHMKIHLHOMIHI/L;5L;5h I9 Lf.HF0HD$8HE(HD$0HE LHD$(HEHD$ GIIhI~aIu&M~jH57LHVcNHtHD$8IM~D1HL$ MLL H5Š萠y kfMtfIuMHD$0Ll$ Ld$(HD$HD$8H$ f.H|$覛L|$IfLFIH5^LIHVMHD$ H{LmMH֭H ǭHNAL ǷH.ELNOD@HH H AUH5H81PXZH fH=E1 @H MAL OfQIWLG[LGAH$H5ŗHH4HIH( H;H;W,L;D$!LLD$$MLD$AL I(j EH5bH<$GIHH;AH;ޖD I9 LLD$LLD$AI(U E HH H9HrLMII@LD$LH5rHHLD$IMI( IGH52LHHIMI/ H<$LƺLD$FLD$HII( L; AL; D I9v LLL$KLL$AII) H $HHD$HHEfDHQH H9HxLMII@LD$LH5HHLD$IIHMIH;H|I9GLLoIIHMIHxIALL$LH5HHLL$IIHMIHH5LǺLD$DLD$HII(L; DL; OI9FLLL$ILL$|I)xH$H mHH9H H<$L[IH$HHD$HM H $HHHm.HLmCH^LXCLD$FCLD$D@L0CHD$HE1LLLLH IAWHT$}IXZMg I/jLB]LHAƅAlI/ H DH=DE1Af.I/LD $ZBD $pHFLHD$ @IDH5LHV-GHtHD$(IM H5gLHVGHHD$0Iq@H=HH5 IM:LMIzADHhA1LvIIHDE1A<DH= Iq@[JIL@{A?pL@L@L@L@H $HHD$HH'H<$~@ fH=QHbH5cv IMMLLL$+@LL$qH= ISII!E1ADLLD$?LD$bfDfLHD$?LL$ffHHNH ڤH=`SfDLP?cMPCAHHH5D $H8]@D $@AHH H9H%L=MJIIGH5)LHHIIIHM$IHFLH%?IHI.L;=pL;=>I9LDAƅXI/EHD$HE1LLLLH IAWHT$IAZA[IHAQL=E1E1MAN}H=yHbH5c IMeEAg{FIH= IL=LH<\E1AjDH=HH5IMvAf.L<ELD$IgfDH=QIAUL8<AMGMIIGH5LHHIIHMIHLH8IHI.L;=ZL;=(I9LLD$IPH=4IAMAML4$P>ICApD8MMAL4$H H H9H"L5MIH5<LdIHLpLH5IH'LLLDADžL1EjHD$E1LLLLH IASAVHT$IXZMLME1AH=IMaH=HH5IL3t.=I4AQL3L3^AE1E1wAH=IM<GH=HwH5xIACE1E1E1APAnAAE1Al`AiH=SIMPgH=,HH5IIMǻH<$DL$LD$0LD$DL$,H=H9tdHEt H;MuNHk9IMtqH=LTIHtJL4L<|L$H3IAE1E1HAAE1AH= IMKH=H[H5\If.AWAVAUATIUSHHF:LnIEHHHcH>ID$H5TLHHIMs4IHH1HHH5HL5HmHH ;H9HH-"HYHEHEH5HHHfIM@HmH5~LL;5I.YIGH-NHH6H=R,jLHLIw2MfI/$I,$ IFH5LHH HHI.3Ht|H9]*L}MLeII$HmUH~LL;{I/INMI,$IFH5?LHHIIHIMHI9_MwMMgII$I/ LL|I.HHlI,$HEHXpH(H{L.IH%H={HHo4I,$IMLHSImDHUHJHHMHfH[]A\A]A^A_DnLImHHIHHDnFII DnFII IIu/IHgH 9H=1eDnAMc@HPH}zH5ގH81<3HUHBA$HEHDH H=yDH1[]A\A]A^A_A$HuL(*fDH*A#E1I/hI,$HtHmtYMlI.bH.I@H)H5yHI6zIfDH)fDHp)I,$A$1E1DL@)NLM|H0DL)L)L(A#1fDA#\ 2I_L(I/uL(A#fLx(E1A#[DHP(L@(L0(JLHL:0IHRfL'HHD$'HD$fH=HH5HHE1A$L'H=iHA$V0IL8'LHD$#'HD$fL'A$;0HA#E11s+HH1A#DL&/I,fDH@`H`HHPHHH?H@H;vLoIEHHHcH>DoH/%H/Y%E1@DoAMcH/f.DoGII IDDoGII yH5 HHdH@$H|$*H|$IH|$&H|$IsHtA#H5H8X&)HHtH5֊H8/&f.AWAVAUATIUSHHtLnH|$HdH%(H$1HtHDŽ$H$H$H$H+I II,HNtH$HD$8ID$ HD$Ml$HD$HDŽ$HDŽ$H=2HDŽ$IEHH$HHH9x,H-H.HEH$HEH5HHH-IM-Hm'H$HHr.IELh)H$IH.H!sH5H*ID$LM,H=!11LHLAIHD$0(M1HD$0H$I,$W)Hm<)HDŽ$I/)HDŽ$Im(H|$0H5#HGHH?1HH$H/H5H9tHEH;zr4H}H2rAHIL$Hm*HDŽ$I/*HDŽ$EO HD$H@HL8I9t MHPHuLhLpMMtIEMtIHhH H9H/8H-xH%<HEH$HEH5HHH<IL$M;Hm0H|$0H5HDŽ$HGHH;HH;HZpHD$xH9G,HoH,LWHEIH/6LHLT$qLT$H$IHmD-M<I*2HD$xI9D$}?H$LLL$pLL$H$HI)r2HDŽ$H$H?H/#6HDŽ$HDŽ$MtI/uLMtImu L@MtI.u LDH5IH, H$IHB?H;yoL;=Go+I9+L%Aą@I/5HDŽ$E HH H9HOSL=MXUIL$IGH5ZLHH!UIL$MvTI/9HDŽ$HD$xI9FVHt$H$oH$IHDŽ$MUH$H/9H5L94IGH;mn^I4H%nAHIL$I/h<HDŽ$I.=HDŽ$EL-LL%5IEHHpH=J~1LLI"M؈L$LjI/hHD$pE1E1H$HD$8E1E1Ll$0HDŽ$AE1HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$,@I~$HF0H$ID$(HD$8bHHlH$H\$HD$8;ff.Ez ,DH1lE1L='lHHDlH9lA@HD$pE1H$HD$8E1E11HD$hAE1HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$Ht H/1Mt I,$YH$Ht H/zH$Ht H/Mt I)Mt I(}H wDE1H=oHt Hm2Mt I/Mt I.HL$HtHHD$0HHH\$HtHHD$HHHL$(HtHHD$HHwHT$ HtHHD$HHhHt$pHtHHD$HHYH\$@HtHHD$HHJHL$PHtHHD$HH{HT$XHtHHD$HHlHt$HHtHHD$HH]H\$`HtHHD$HHNHL$hHtHHD$HHHT$8HtHHD$HHImH\$HHD$HHH4$HtHHD$HHtqH$dH+%(HL[]A\A]A^A_DLhLpIUID$(H$HD$8DL^HfDHH#HxxHhHXHHL8PL(OH^HmH|HHHHH|$fL$D\$xLD$0D\$xLD$0L$7fLD\$xLD$0^D\$xLD$0#L$D\$xLD$01D\$xLD$0L$f.LL$D\$xLD$0D\$xLD$0L$vL$D\$xLD$0D\$xLD$0L$Xf.LD\$0D\$0lfH|$0H5$HGHH0HH$H'0H5H9#HEH;wf2H}u}#H!fAHIL$Hm$HDŽ$I/(HDŽ$E 1H|$0H5HGHH14IL$M-4IGH;f+H; er"H@hH6H@H51LHH$HUEI/+H9-HDŽ$HDŽ$$H$H9"HqIHEHH H9HdFL=MIIL$IGH5LHHRJIL$MII/;/HcH LHDŽ$H9HJL='MKIIGH5LHHKIMLI/d0HH H9HOL=MQIIGLL$LH5HHRLL$HD$H|$HRI/4H$cHD$xI9AUHt$LLL$ dLL$H$IHL$HHD$HH3MUI) 4IGH5kLHHWHD$H|$eVI/CHDŽ$HD$xI9F[Ht$H$fcH$IHL$HDŽ$HHD$HHZCMXH$H/YCH$HHDŽ$HDŽ$H@H9t:HXH\HqH1DHH9qH;TuHHH9PeL5M=iIL$IFH5 LHHjIMIIFH5LHH‘IM"I.9nHHHH9PٛL5MIL$IFLD$LH5hHH͚LD$HD$HD$H$HI.mH<$H5LD$HGHHLD$IL$MHD$HT$xH9PH$LLD$ _LD$II.:wHDŽ$MԎH$H/vIALD$ LLL$H5cHHOLL$LD$ HD$HD$H$HI)yHD$xI9@(Ht$LLD$Y^LD$H$IHT$HHD$HHxHDŽ$MI(ILL IHH;a]AH;.]DtI9tLLL$LL$A I)1ExILH$I/։I.L$I<$ƉHDŽ$I/MH L(hH<$E11HA1AH$IH-YH8CHDŽ$H $HHD$HHCHDŽ$I91[IFHMfLH5H\HD$HD$H$H \H5TH;t$FHT$HBH; \meHzuHD$xFH[HHD$HD$HL$H$HH$HHVHHDŽ$HL$HH$HH.IHDŽ$fIFH5LHHUbHD$HD$H$H_eHH HD$H$H8oHT$HH$HHQHL$H; ZHDŽ$H; Z AH9AHYiHL$HH$HHSHDŽ$kjLLPH)H¿$H9PxHHD$HvHH$H|$H5HGHHvIMnyHT$HHD$HHbV$LD$ LD$HHD$H$tHD$xI9@I}Ht$LLD$ ZLD$ HD$H$HT$HHD$ HHUH|$HDŽ$I(VHL$H; %YH; XxNH9oNHAŅHT$HHD$HH\HDŽ$E~HH 'H9Hx~LM}II@LD$LH5nHH}LD$HD$HD$H$H|I(oH|$H5HGHHIM\HL$HHD$HHmH51LLD$LD$HHD$H$HD$xI9@\M`M\MhI$IEI(>tHT$LLVH$I,$HD$@yHT$HHD$ HH mH|$HDŽ$ImyrHL$H; ,WAH; VDkTH9bTH AŅHL$HHD$HHqHDŽ$E$\mfTmx HD$H$HH|$L IHHT$HH$HHexL;\VHDŽ$AL;VDljI9cjLL$ L$A4I(wL4$E*fDH9\$ HUAHIHL%+HL$HHD$I$HHD$HH*8Ld$HL$8H; UH; `U-H9$H4 H&H<$H97HGH5HHjRHD$HT$H$HUHTHD$xH9BD7LbM77HzI$HHH$HD$ HHHLJUH$I,$HD$((OH|$(H$WH/5BHD$(HDŽ$H@H;TEH;SNL`hM>_I|$2_I$HkH|$(HHaHH|$(AT$HD$ HD$ H$HRYLd$ H|$(LHD$H$H-[I$HD$8HI$AHHT$(HHD$ HH6HH|$HH5HDŽ$HGHHgHD$(HD$(H$HniHT$(HD$xH9B?LbM?HzI$HHH$HD$ HH[MHT$LQH$I,$HD$(2PH|$(cH$H/^KH|$H5wHDŽ$HGHH>aHD$@HD$@H$H`HD$PH$HjHD$(HT$PHHBIHjHH5ٲHHD$ LD$ vSHD$@H@LM_H=bLD$ LD$ l{LD$8LHt$PH|$@AIHD$ MLD$8zHD$ H$HT$@HHD$8HHRHL$PHDŽ$HHD$8HHRHDŽ$I(qRH&HHDŽ$H9PyHZHD$@HYyHH$H|$@H5ܲHGHHyIMxHT$@HHD$8HHYLD$8LD$8HHD$PH$NHD$ HL$PLD$8HHALD$8HHD$XH$H0PH|$XLD$8H5LD$8BjHT$XHt$PLLD$8LD$8HHD$@H$}I(iHL$PHHD$8HHiHT$XHDŽ$HHD$8HHiH|$@H5uHGHHۓHD$HHD$HH$HHL$@HHD$8HHwpHD$@HH$HHhOHHBHD$PH$HHγH5H|$PǐHT$PHt$@H|$HHD$XHUH|$HAH|$@HDŽ$+H|$PHDŽ$H|$ HDŽ$HD$XHD$pHD$hHD$ HD$`HD$HHD$XHD$PHD$@EJHD$ HQHxH9t5HXH|HqH1 HH9H;TuHH5]H|$ HHD$8HD$8H$HOHT$8H&MH9B7eLbM*eHBI$HH$HHD$xHHZvH$HLKLHD$8H$豪H|$8H$H/dHT$ HDŽ$HHD$xHHhlHD$8HD$ H|$0H5wHGHH!qIL$MpH5VL9HID$H;LyI|$GHXLHHD$8HD$8H$I,$UHDŽ$1HD$8H;@LHL$8HHD$xHHcHDŽ$JEoaH5H|$ MH$IHșH5H9H@H;KI|$HKHHD$8HD$8H$LH|$8HDŽ$ÅH|$8ݨHDŽ$`H*H3H9XJHHD$8HHH$H5H|$8_H$IH%H|$8HD$x\GLL$xHH$I^H5H|$0LL$8LL$8HIUH5hHLLL$xHD$8LD$8LL$xLLL$8ܧLL$8H5LLLL$xLL$xHHD$8L襧LHDŽ$葧Ht$ H|$0HDŽ$6QIH3H5H|$8HHD$xLD$xӪLfHF0H$ID$(H$ID$ HH$ID$H$5III~jIu)M~wH5HHVHtH$IM~N1H$MHL ^H5P Py'yDMIuMH$L$HD$H$HD$8H$H$DH ]AL gHHGH^H5YAUH81XZH B]'H==\E1HIH5HIHVH$H$Ml$MH\H \AHNL fEHFLNOD@;H@HDŽ$AL 7H=H:H5;HH$H;HD$pE1E1E1HD$8E1AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$fLHLrHD$0H$HzHD$pE1E11HD$8E1H$AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$%DIHD$pE1E11HD$8E1H$E1HD$hAHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$~fDH=H@HD$pE1E1E1HD$8E1AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$HD$pE1E11HD$8E1H$AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$IH|$FLT$H$IMDwHFHH$IH5HHVHtH$IMH5HHVdHdH$INf.HLL$ LT$LL$ LT$IoHEH$HH9HuH;CYf.EHBE1L=wBHHBH9eBAp@E1HD$pE1E1H$HD$8E1E1Ll$0HD$hE1AHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$@HD$pE1H$HD$8E1E11HDŽ$AE1HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$fH|$0fHfDHHDŽ$HH*HD$pE1E11HD$8E1H$E1HD$hAHDŽ$HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$@HH"H9PH*L5 MQ+IL$IFH5qLHH+IL$M%+I.vHDŽ$H;?I9G-Ht$H$'@H$IHDŽ$H$M,H/H5L9 IGH;?4I H:?AHIL$I/HDŽ$I.HDŽ$EL-aL%2IEHH<H=_OR1LLIMWL$L;I/7HD$pLl$0A-HDŽ$HD$8E1E1ɻHD$hE1E1HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$k@LLLL$LL$XfLH;1=%H&H$IHHD$pLl$0E1E1HD$8E1H$E1HD$h1AHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$CH<E1HH<H9<AHD$HHHH9P+HHD$H-HH|$H5HGHH,HD$(HD$(H$H2,HL$HHD$HH 7IH/HD$LD$HI@dLD$HHD$H$2HH#H9Pf4H HD$ H6HH|$ LD$H5gHGHH6LD$HD$@HD$@H$H5HT$ HHD$HHHT$@H|$LD$H5 LD$$HT$@HHD$HHD"HD$(H@LM/H=KLD$LD$B8LD$LHT$H|$(AIHD$ MLD$'?HD$ H$HL$(HHD$HH#HDŽ$I(#HT$HHD$HH#HL$HDŽ$HHD$HHt#HD$L59HDŽ$HD$HD$ HD$_DLT$LT$=@fDIGH(HELhL]ffA.G_YDHQ9E1L5G9HHd9H959A<@H=њHBH5C6HH$HH$A Ht/H/\D\$D\$HDŽ$f.H$Ht H//H$HDŽ$Ht H/ Hl$H5zD\$HDŽ$H}X[BD\$H MDH=L#H$H$HH$ H-L%HEHH$H=H$1LHIMl3L绋M4I,$Aa$@HD$HH8L`L8HhLhLpHt H/Mt I,$HHmHD$pLl$0E1E1HD$8E1E11HD$hE1H$HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$,@HD$pE1E1H$HD$8E1E1Ll$0HD$hE1HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$HD$pE1E1H$HD$8E1E1Ll$0HD$hE1AHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$H|$0HffA.G@:HL5E1L5B5HH_5H905A"L/H=˖辰HA KVIAQVLBH|$0'HH$ArH-;L%HEHH(.H=9Eg61HLHH:H$H0Hm!HD$pLl$0E1AHD$8E1H$HD$hE1E11HDŽ$E1HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$LLL$LL$H$HtH/fHDŽ$I*ALD\$ID\$H;2zHH$IHHD$pLl$0E1E1HD$8E1H$E1HD$h1AHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$D\$pD\$D\$\D\$D\$HD\$LD\$1D\$HD\$xE1E1E1Ll$0E1HD$pD\$xE11HD$8H$HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$L~HDŽ$H|$0IHD$pLl$0E1E1HD$8E1H$E1HD$h1AHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$HH ]H9HPHDHD$HOHH$H|$H5HGHHOHD$ HD$ H$HNHT$HHD$(HH1H5LH<$.HD$(H$HbSH+HT$ HD$xH9BSHt$(H$,HD$H$HL$(HHD$ HHy3H|$HDŽ$VH$H/8Ht$H|$1HD$(H$HZHT$HHD$ HH8HL$(H; +HDŽ$H; w+ (H9(HKAąjaHL$(HHD$HH%8HDŽ$EnHH<$H5׌HHD$pHGHHZHD$HD$H$HYHT$HD$xH9BDLbL$MDHBI$HH$HHD$ HHADH$L"+LHD$ H$H|$ HDŽ$YH$H/7DHDŽ$H$HHD$HHDHH HHDŽ$H9HoH#HD$HCoHH$H|$H5eHGHHoHD$PHD$PH$HiHL$HH$HHMHD$(H$H1]HD$HT$(HHBHD$H$H^Hw)H5H$^HT$Ht$(H|$PԥHD$@H rH|$P|H|$(HDŽ$fH|$HDŽ$PH5IH|$@HD$H$H@qHL$xH9Hu3H@H|$H$H_HWHHH$L$H$M^L(LHD$P轅H|$PH$HDŽ$Ep蘅H,$1E1HDŽ$HD$hHD$`HD$L$Dl$(L$IHt$H|$p1_HHpoH+AŅyHEyH5H|$HޢH$IHyHt$pH|$KH$IHxHDŽ$HD$xI9GZIGH$HZIWHLHH${H$HZH$LHg%HHD$XJHDŽ$L6H|$XHDŽ$UH$HDŽ$MtLH5!H|$pHH9wHΨAƅvH軃EDHH<H9P L=#MIL$H5LOH$IH~LSHD$xHDŽ$I9En~Ht$ H$6&IHDŽ$M}H$HD$HDŽ$HtH1ɺHL`)HH;}HLHD$H$H|H蛂L蓂H5̅H|$bH$IH|HH{HD$XHHE6H$IH({HcH5TH<zLHLFH$IH"zLHHDŽ$ցL΁HDŽ$MtL赁HH /HDŽ$H9HyL- M yIEL$H5]LEH$IHxLIH$IHILpHH H#H5HHLL&H$IHLʀLHDŽ$趀HHDŽ$袀ID$H;"؅I|$HPIl$ID$ H$LHEH^HDŽ$HD$`HtH@H$HD$`HD$hHtH!H5H|$`H$IHڋHD$xI9GMgMIGI$LHH$H$L"LH$IMH$LH5SLHDŽ$GH$IH1HD$xI9GMoMIGIELHH$H$HT$`L LH$I~M5H$~LHDŽ$~H5L虜H$IHH|$pHxH$HL~HD$PHDŽ$H@LppMSkI~HkH$H|$pH}IHjH|$PLHAVI/AjEjH$}H5LЛH$IHkH|$pHH$IHPkL}H|$pHDŽ$}HD$`Ll$pHl$`Lt$XHDŽ$HD$h#HD$pE1E1H$HD$8E1E1Ll$0HD$hE1AHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$L@HH3陼L&4靼HD$pE1E1H$HD$8E1E1Ll$0HD$hALE1HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$HLD$rLD$1H=QHJH5KnIL$MHD$pE1E1H$HD$8E1Ll$0E1HD$hAVHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$H|$ HD$(H$HDŽ$LD\$kD\$HYLd$H|$HH5+HGHHj>IM=LD$BLD$HHD$(H$/=HHT$(LD$ HHBHEHj PLD$ HHD$HH$=HT$H5QLD$ (HT$Ht$(LLD$8MLD$8HHD$ H$GI(*)HT$(HHD$8HH}(HL$HDŽ$HHD$(HHr'HDŽ$HDŽ$HD$pHD$hHD$`HD$HHD$XHD$PHD$@HD$(HD$H=HH5谗IL$MHD$pE1E1Ll$0HD$8E1AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$鉯DH=}IHD$pE1E1H$HD$8E1Ll$0E1HD$hAXHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$鸮I馵HD$pE1E1H$HD$8E1Ll$0E1HD$hAHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$IתH=z~%ID锹H=U~H>H5?rIM HD$pE1E1H$HD$8E1Ll$0E1HD$hA[HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$)HLD$LD$HVH$I\H=\}IHD$pE1E1Ll$0HD$8E1AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$lMfL$MIFH|$H9HXH62HqH~1H;|HH9uH\$HPH5E1HaHKH81L4$E1E1HD$pLl$0E1AIHD$8H$HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$UDL4$Ll$0E1E1HD$pE1H$E1HD$8ASHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$蓸HD$@HD$pE1Ll$0E1HD$8A HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$([L輮邭HLD$8誮LD$8UHLD$8蓮LD$8HD$pE1E1H$HD$8Ll$0A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ 鍒HD$pLl$0E1AHD$8E1H$HD$hE1E11HDŽ$E1HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$HD$pE1E1Ll$0HD$8E1A HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$o1LL H$IHlHD$pE1E1H$HD$8E1Ll$0E1HD$hAHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$鹐H9rH=aHaH5a:xIL$M&HD$pE1E1H$HD$8Ll$0E1AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$HlLD$HR9HD$Ll$0E1E1HD$pL5$A| HDŽ$H$HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$(HD$HD$2Ht$LLD$ HLl$ HD$H$4HD$(Ll$0E1E1HD$pE1H$A HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$HD$(鈎HD$pE1E1H$HD$8Ll$0E1AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$LLD$zLD$H$(HD$pE1E1H$HD$8E1E1Ll$0HDŽ$E1AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$6 HD$飝H=]CtIH蓨zHD$(Ll$0E1E1HD$pA HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$HD$(馌HD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HLD$8肧LD$8ЦHD$pE1E1H$HD$8Ll$0E1AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$^1IHL$8H; H;  H9 I,$uL襦HDŽ$[H|$8gÅXLl$0E1E1HD$8H$A$E1ƊH;H߻!ML4$E1E1HDŽ$Ll$0E1AHD$pH$HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$lHD$pE1E1H$HD$8Ll$0E1A HD$hHD$`HD$HHD$XHD$PHD$@HD$ H5NH|$HiHD$H$H8XH|$H H9OLgMHGI$HH$KH$HLLHD$H$KH|$4OH$hKHD$H@L`pM"I|$"Ht$HHvH"HD$(H|$HAT$LD$(HD$ I(v"HD$ H$HHH|$JH|$ HT$HDŽ$HDŽ$H5OHGHHLHЅHD$pE1E1H$HD$8Ll$0E1A#HD$hHD$`HD$HHD$XHD$PHD$@HD$(HD$HD$pLl$0A)HDŽ$1LLLD$LL$tLL$LD$HH$I HD$pLl$0AHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$LE鯕H|$HHD$(AL4$Ll$0E1E1HD$pE1H$E1HD$8A{HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$~Ll$0Ht$ LHD$8IHVH$E1E1A$~HLD$-LD$9HӒIAH$HMaHLAI$vGI|$MA釈HD$pE1E1H$HD$8Ll$0E1A HD$hHD$`HD$HHD$XHD$PHD$@HD$ }H5 JH=M1fHD$H$HK6H\$E1HH߻FLl$0E1E1HD$pA!H$HDŽ$HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$|HLL$zLL$zLh1ENHRBH5KH|$8HD$8H$ %HD$pE1H$E1HD$8Ll$0A !HD$hHD$`HD$HHD$XHD$PHD$@HD$ $|HD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$@HD$ {LL$LD$/LL$LD$LLL$LD$LL$LD$饈誛H9HDŽ$GHD$pE1H$E1HD$8Ll$0AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$zHIL$MH$HD$ H$鬻H鿻HTH!LHD$pE1E1H$HD$8Ll$0A+!HD$hHD$`HD$HHD$XHD$PHD$@yL4$E1Ll$0E1HD$pAHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$yHILHLD$ȔLD$/LLD$豔LD$͆L蟔zEH8HD$pE1E1H$HD$8E1Ll$0E1HD$hAHDŽ$HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$cxHD$pE1E1H$HD$8E1E1Ll$0HDŽ$AE1HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$wL?鵋H|$耜ISH=H^HD$HD$H$HL4$Ll$0E1E1HD$pE1AHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$wHD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$kvH5BH=uF1^HD$H$HhAH\$E1HXH߻;?Ll$0E1E1HD$pA!H$HDŽ$HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$uH=FHFH5F$]HD$HD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(tUHmHD$8HD$ 鋓L4$Ll$0E1E1HD$pE1H$AHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$@tHD$pE1H$E1HD$8Ll$0AX!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$sHD$pE1H$E1HD$8Ll$0AV!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$=sH|$H IHD$pLl$0E1A`!HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(rLLL$4LL$L"*L鳆Ll$0AHDŽ$HD$pHL$ڍL$醇Lɍ€L輍GHD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$PHD$@HD$(HD$qH|$}HD$ `H= BXHD$HD$H$HHD$pE1Ll$0E1HD$8A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(qH=pAHAH5AXHD$QLL$LD$LD$LL$H/HDŽ$M`MMhI$LIE9HT$LLMH$I,$HD$LLl$ ՋLD$ vHË;Ll$0E1E1ɻHD$8H$A $ pH|$0ؔIҎH|$HHD$H$pL4$E1Ll$0E1HD$pAHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$vo*LD$HD$yH=?_VIM2L4$Ll$0E1E1HD$pE1H$AHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$nH=?H?H5?#VI?H5:H=}>1VHD$H$H.H\$E1H`H߻C7L4$E1E1HDŽ$Ll$0E1AHD$pH$HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$mLM \|HD$pE1Ll$0E1HD$8A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(mLbHM4HBI$HH$5HT$(H$LH$I,$HD$ L2H%ֈHD$pE1H$E1HD$8Ll$0A!!HD$hHD$`HD$HHD$XHD$PHD$@-lH|$@I؆H=<7SHD$@HD$@H$HHD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$PkH=<HY<H5Z<SHD$@aH陉葋LD$8HC'HD$pE1H$E1HD$8Ll$0A!HDŽ$HD$hHD$`HD$HHD$XHD$PHD$@jHD$pE1E1H$HD$8Ll$0A!HDŽ$HD$hHD$`HD$HHD$XHD$PHD$@HD$ TjH=:H~;H5;QIM)vHD$pE1E1H$HD$8Ll$0E1AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$iHD$pE1Ll$0E1HD$8A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(FiHD$pE1H$E1HD$8Ll$0Ac!HD$hHD$`HD$HHD$XHD$PHD$@HD$(HD$hL4$Ll$0E1E1HD$pE1E1H$HD$8AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$hL4$E1Ll$0E1HD$pAHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$gL4vLvL-vH;\L膃HD$8H$HLl$0E1E1E1H$A$gL4$ME1Ll$0HD$pE1AHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$fHD$pLl$0E1A,!HD$8HD$hHD$`HD$HHD$XHD$PUfHD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$eHH5H8xHHH9=HuH;D+&HD$8E1E1H$HD$hLl$0E1A"HD$`HD$HHD$XHD$PHD$@HD$ HD$(dH<$ɉHD$tHD$8E1Ll$0E1HD$hA"HD$`HD$HHD$XHD$PHD$@HD$(HD$dHD$pMH$AHD$8E1E1Ll$0HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$cHD$pE1H$E1HD$8Ll$0AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$ZcL`HMQpH@I$LD$HH$C,H$LL@LHD$#,LL$LD$ pHHH9T$QrHuHT$H;:rH=<3IIMLl$0E1E1HD$pAH$HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$%bI6nHD$pE1H$E1HD$8Ll$0AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$aH<$XLD$IUnHD$pE1E1H$HD$8Ll$0E1AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$`Ll$0E1E1E1H$A#`H$LHD$XxHD$ Ll$0E1E1HD$8E1H$A1"HD$hHD$`HD$HHD$XHD$PHD$@HD$H$HD$ *`HPHH5H81FHD$pE1E1H$HDŽ$Ll$0E1A#HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$y_Lz}HD$ Ll$0E1E1HD$8A8"E1H$H$HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$^HD$ Ll$0E1E1HD$8A6"HD$hHD$`HD$HHD$XHD$PHD$@HD$(H$HD$ ^HD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$]XHD$P1L4$Ll$0E1E1HD$pE1H$A HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$@]L4$Ll$0E1E1HD$pE1H$E1HD$8A HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$\L4$E1Ll$0E1HD$pAHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$M\H5(H=8,1qDHD$H$HZ)H\$E1HH߻$L4$E1E1HDŽ$Ll$0E1A HD$pH$HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$I[HD$pE1H$E1HD$8Ll$0A HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$ZHD$pE1H$E1HD$8Ll$0AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$4ZLD$ LL$HD$gM`MgIHI$LHHL$ #HT$H|$L LH$I"LD$gHD$pLl$0E1AHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$`YL4$Ll$0E1E1HD$pA E1H$HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$X~}LD$HD$+eH=)LD$?LD$IL$MdHD$pE1H$E1HD$8Ll$0AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$WH=O(H(LD$H5(g?LD$I4HD$pME1H$HD$8Ll$0E1AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$ WH1H5E1仟H8sL4$E1E1HD$pLl$0E1AIHD$8H$HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$zVHH5H8%1w=HD$HH\$E1E1H&H߻ Ll$0E1E1HD$pA, H$HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$gTHD$ Ll$0E1E1HD$8E1H$A."HD$hHD$`HD$HHD$XHD$@HD$(HD$H$HD$ SHD$pE1E1H$HD$8Ll$0E1A1!HD$hHD$`HD$XHD$PHD$@SHD$pE1E1H$HD$8Ll$0E1A9!HD$hHD$`HD$HHD$XHD$@SHD$pE1Ll$0E1HD$8A4!HD$hHD$`HD$HHD$XHD$PRLl$0E1E1E1H$A#RpwHD$8fpHD$pE1E1H$HD$8Ll$0E1A;!HD$hHD$`HD$HHD$XHD$PHD$@ RL4$Ll$0E1E1HD$pE1H$AHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$QH|$XvIdL4$Ll$0E1E1HD$pE1AcHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$QH5H= 169HD$H$HH\$E1HH߻Ll$0E1E1HD$pA!H$HDŽ$HD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$PH|$@tHD$HlHD$pE1H$E1HD$8Ll$0A)!HD$hHD$`HD$HHD$PHD$@OHD$pE1H$E1HD$8Ll$0A$!HD$hHD$`HD$HHD$XHD$@FOffA.D$!ooHȺHHD$8oH=66HD$HD$H$HHD$ Ll$0E1E1HD$8E1H$A,"HD$hHD$`HD$HHD$XHD$PHD$@HD$(H$HD$ sNH=HH56HD$FH|$sHD$PHD$pE1E1H$HD$8E1Ll$0E1HD$hA( HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$MHD$pLl$0E1AHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$JMHD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(LILLd$HH,$HD$XHD$ E1E1HD$8L$Ad"HD$(L$H$H$HD$ Ll$0@LHD$ Ll$0E1E1HD$8E1AV"HD$hHD$`HD$HHD$XHD$(HD$H$HD$ KHD$ Ll$0E1E1HD$8AH"HD$hHD$`HD$HHD$XHD$PHD$(H$HD$ KHD$ Ll$0E1E1HD$8E1H$A9"HD$hHD$`HD$HHD$XHD$PHD$(HD$H$HD$ JHD$`ILd$HE1LD$`H,$A{#HD$hHD$ E1E1HD$8L$H$L$HD$ H$HD$(Ll$0pJLefHD$(HD$`IE1Ld$HH,$E1A{#HD$hHD$ LD$`Ll$0E1H$L$L$H$HD$8HD$ IHPHl$(H H>H57H,$Ld$HL$L$E1H81mHD$`LD$(E1HD$8Ll$0A{#HD$hHD$ LD$`E1H$H$HD$ HD$(SIHD$8HD$`IE1Ld$HH,$E1A#HD$hHD$ LD$`Ll$0E1H$L$L$H$HD$(HD$ HHD$(HD$`IE1Ld$HH,$E1A#HD$hHD$ LD$`Ll$0E1H$L$L$H$HD$8HD$ qHH; HH;ҳ H9 HL$$HHD$ HHuHωT$ cT$ HDŽ$<$5XH|$ii…!XL4$Ll$0E1E1HD$pE1E1H$HD$8AXHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$^GHD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(FHD$pE1E1H$HD$8Ll$0E1ALibAŅ>HD$pE1E1H$HD$8E1E1Ll$0HD$hAE1HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$^@HoH5H8 ]?L4$Ll$0E1E1HD$pE1H$E1HD$8AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$?HD$(HD$ ME1LL$HH,$E1A"H$L$HD$8L$HD$ Ll$0{?H=x&IL$MHD$ MLl$8E1LL$HH,$E1E1H$L$A"HD$(L$HD$ H$Ll$0>H=KHdH5eh&IkIHD$ Ld$HIHD$8H,$E1A"H$L$HD$(L$HD$ H$Ll$0U>HD$ Ld$HIE1H,$Ll$0E1HD$8H$L$A"L$H$HD$ HD$(=HD$(HD$ IE1H,$Ld$HA"HD$8L$H$L$HD$ Ll$0=IHD$ Ld$HE1H,$Ll$0HD$8E1H$L$A"L$H$HD$(HD$ ;=HD$8HD$ E1E1Ld$HH,$E1A"H$L$HD$(L$HD$ H$Ll$0SHHD$ LLd$HE1H\$XH,$E1E1H$Ll$0A"L$L$HD$8HD$(H$HD$ F9H|$ TY駷HD$pE1E1H$HD$8Ll$0E1A#HD$hHD$`HD$HHD$XHD$PHD$@HD$(HD$8HD$ MILl$0LL$HH,$E1A#HD$8L$H$L$HD$(H$HD$ T8HD$ MILl$0LL$HH,$E1A#H$E1L$L$H$HD$8HD$ HD$(7IHD$ MH,$LL$HE1L$E1H$Ll$0A"L$H$HD$8HD$(HD$ 7HD$8HD$ ME1LL$HH,$E1E1H$L$A"HD$(L$HD$ H$Ll$0&7H;AL;=AEA I9A HmuHvRHDŽ$Es0L9XAŅ`0HD$pLl$0E1AHD$8E1H$HD$hE1E11HD$`E1HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$,6HMHH8S/UH~ML4$Ll$0E1E1HD$pE1H$E1HD$8AHD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$n5HD$(Ll$0E1E1HDŽ$A E1H$HD$HD$pHD$8HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(4HD$pE1E1H$HD$8Ll$0E1A!HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(X4H;AL;5AEA I9A I/HDŽ$E`LpUAŅ`HD$pE1E1H$HD$8E1E1Ll$0HD$hAE1HD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$HD$e3L5֞ ILl$0E1E1H$Ae$63LNHD$pE1E1H$HD$8Ll$0E1A#HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(2L4$Ll$0E1E1HD$pE1H$E1HD$8A HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(HD$2MH,$L$L$HYHHD$ LL$HE1E1H$Ll$0HD$8H$A #E1HD$ HD$(1H;TLcTH$IHLIGLHDŽ$LAHH#LAH$HjLAԾH"L$HDŽ$yL $E1"!L $HD$ E1HD$8H$A #LL$HLl$0E1H$HD$ HD$(0HL $E1H5H81TL $HD$ E1HD$8H$A #LL$HLl$0E1H$HD$ HD$(+0HD$ MILl$0LL$HH,$E1A*#H$E1L$L$H$HD$8HD$ HD$(/L|$(IMH,$L$L$AH|$(LL$8L$HDŽ$L$LL$8tsHD$ LL$HA2#HD$8Ll$0E1E1H$H$HD$ HD$(#/L|$(MH,$IL$L$E1\LL $E1LD$xL $HD$ HD$8Ll$0LD$xHD$ A2#LL$HH$E1H$HD$(.HD$(HD$ ME1LL$HH,$E1A"#H$L$HD$8L$HD$ Ll$0^.I|$HID$H(H@vLl$0E1E1H$Ac$-Ll$0E1E1E1H$AU$-Ll$0E1E1HD$8H$AS$-Ll$0E1E1HD$8H$AQ$~-HD$8HD$`IE1Ld$HH,$E1Ax#HD$hHD$ LD$`Ll$0E1H$L$L$H$HD$(HD$ -HD$(HD$`IE1Ld$HH,$E1Av#HD$hHD$ LD$`Ll$0E1H$L$L$H$HD$8HD$ ,HD$`IMH,$LD$`Ll$0E1Ai#HD$hHD$ LL$HE1L$H$L$HD$8H$HD$(HD$ ',Ht$`LBH$I0vHD$(HD$`IMLD$`H,$E1E1HD$hHD$ A[#LL$HLl$0E1H$L$L$H$HD$8HD$ +HD$`IML|$(LD$`H,$E1E1HD$hHD$ AO#LL$HLl$0E1H$L$L$H$HD$8HD$ "+L蒙H$ItHD$(HD$`IMLD$`H,$E1E1HD$hHD$ AA#LL$HLl$0E1H$L$L$H$HD$8HD$ *HD$pE1E1H$HD$8Ll$0E1A#HD$hHD$`HD$HHD$XHD$PHD$@HD$ HD$(*HD$ Ld$HIE1H,$Ll$0E1HD$8H$L$A"L$H$HD$ HD$()HD$ Ld$HIE1H,$Ll$0HD$8E1H$L$A"L$H$HD$(HD$ P)LD`HD$ Ll$8E1E1Ld$HH,$E1A"H$L$HD$(L$HD$ H$Ll$0(LdDHD$(HD$ E1E1Ld$HH,$E1A"H$L$HD$8L$HD$ H$Ll$0x(HPHՓH56Ld$HH,$L$E1H81L$sLHD$ E1E1HD$8H$A"H$Ll$0HD$ HD$('AVAUIATUSHH^H H9HH-HHEHEH5HHHIHEHHEM H7HI9D$LL肓HI$HHwI$HH;-ҒH;-uiH;-t`HxHAąyUA ArHmuHvBDDE1H ǧH=oJHL[]A\A]A^f.DHmyIEEtxCI}HGHHƨHcH>LA"A}@EIHHL[]A\A]A^HH5LHHHHIHHHH59HIyHEL-HHH=2@LLHIWFMHmI,$1L@$A ArH]PfDH@H@zH=H H5  HHBA ArIIFH=I HMt$MMI\$IHI,$LLHI豎I.H%L?H?A HmAuI,$HL?;A}AEHH @A}AEHH HHH|$"DH|$HA AsA}HcDL8?LLHBGIHfDA +LCHoDA AuiL HB HH"A At)A Au+CHbHH54H8?GA Ar}H>>pfAWAVAUATUSHHxH-$LfdH%(HD$h1Hl$`HMILnHD$HHD$PHD$XI9HH H9H_ L%M I$Ld$XID$H5qLHH HHl$PH I,$nAHD$XIH IELhCHD$HIH> HH H9H L%MF I$ID$H5LHH: IM< I,$)H5LLGDWI/5HELM! H=e;- LLHAIAM HmmHD$PI.:HD$XImI<$+ID$At$HD$HI|$ HD$HHD$HLd$1LL=Ml$LLAIHH@HH&LLLIHLL=iMl$LLAAHHH@HHLLHHFH@HD$XH;CLgLd$XM1LI$IH/LLoHD$HII,$HD$XMzI/ImHD$HBH{ IHL$LHZ9IFL-_LMH=C9{ 1LLAIh?M I.I/SHH9l$8Ld$I$LHPI$EMHJH ;HHLH<H?L MLHL@HHATHßH5 H816BXZH ٞLH=E1THD$hdH+%(xHxL[]A\A]A^A_fD|A}E1E1H|$HHt H/H|$PHt H/H|$XHt H/sMt I.tMt I/uH 'DH=MHI$LE1HHH.8$Lx8Lh8IHD$HHD$PHD$XLL=7Ml$LL>IHkH@HHLLLIHLL=Ml$LL=HHFH@HHLLHD$PHHH@H;mLgLd$XMpLOI$ILL$PH/:LL߇HD$HII,$tHD$XMH|$PH/ HD$PIm5IH fDHL8I9t MHPHuLXLPMtIMtIMtIH{ LT$L\$H9L\$LT$HD$HHIzIH8LhL8HhLXLPHt H/jMt ImjHt HmjIFH-HHH=41HLI;MI.I/XL5KHL5L5s5fDL`5LP5~C5+fD35/fDIM<I2HFHHD$`A3HMLl$`fD4Ld$XML轇HD$HIIL4L41LL5HuHH5KH81붐LE1E1]0A+f4IHHH5E1E1~A}H8h11HL&8I.Iu DL/M!E1E1JAz/fDL/H/HLE1E1AzH8/QHLH8/HD$PI.\AzE1E1E1 3HuH~H5H8[0I.uL.JH|$PH F{H= HL$XHT$PLHt$HL\$LT$HD$HHL$XLT$(HT$PL\$ HD$H1HL$HT$8L\$ LT$(HIIFHH:H=LT$8L\$0HL$(LL$ %-LL$ HL$(L\$0LT$8LT$81LLL\$0LL$(HD$ &3LD$ LL$(L\$0LT$8MXI.DI)ML;}L;}I9LLT$0L\$(LD$ P3LD$ L\$(LT$0AI(EIH8LhL8LpLXLPHt H/Mt ImMt>I.E1E1AzLAzE1E1,E1 E1E1AzEHL$HtHHD$ HHHD$HHL$HtHHD$HHHD$PHL$HtHHD$HHHD$XIH8LhL8L`LXLPHt H/Mt ImQIM6I,$+L+I(DLLT$(L\$ +LT$(L\$ RLLT$0L\$(LD$ +LT$0L\$(LD$ LLT$8L\$0LD$(LL$ _+LT$8L\$0LD$(LL$ L>+/4+LE1E1Az+LT$(L\$ G/H\$L\$ HxXLp`HhhLT$(HXXH\$HHX`H\$HXht H/xMt I.Ht HmHD$HHD$PHD$X=E1+LT$0L\$(LL$ /LL$ L\$(HLT$0uHD$(HyH5LT$8H8L\$0t+LT$8L\$0LD$(LL$ BL1LLT$0L\$(LL$ 2LL$ L\$(LT$0I*HLT$ L\$)LT$ L\$BL)HLT$(L\$ )LT$(L\$ HLT$L\$h)LT$L\$T)FLT$L\$@)LT$L\$jLLT$L\$)LT$L\$XHLT$L\$(LT$L\$GLd$E1ALd$E14E1E1AzUDAVAUATUHSHHxLfdH%(HD$1H$H%MIMHH HHLHH?L MLHL@HHwATH H5H810X)ZH K=H=E1HD$dH+%([HL[]A\A]A^fHVHHuHIE1LH= I$LSjSSjSSjSII$HPMt^HI$uLT'yHIMIHFHH$Z%HGH$`@HI$t&H Xx)H=@L&1HL MH5(L0~yx)fH$IHwH5LHV+HtH$IFMB'fAVAUIATUSHH HL%LvdH%(HD$1HMHnH$HD$Ld$MHHHMLNHHHsHH H=sAHEATjQRLj5PjQHIHEHPMHHEtLHD$dH+%(H L[]A\A]A^HLV(HF UfH(%fDIM5DHF M HH fHgIHHI?IAHHtH؋H5XUL ;H81f-X*ZH ~H=E1H AfH^HHIHFLH$}"IM~TH5NLHVB)HtHD$IM~.H5LHV)H HD$IML $HD$LT$H H.HF(LHD$HF HD$HFH$!IHHEt&H ؈D*H=nQ@HP#fDL!IHSH5LHVA(HH$IHF LHD$HFH$/!I1HL IH5!~LIz*#fAWAVAUATUHSHH(L5rLndH%(HD$1H$Lt$HGI5IMHH HNHL LNL@HHqAUHH5H81*X*ZH PH=E1HD$dH+%(H(L[]A\A]A^A_DLV LMHHHHsHHH=AHEHAVjRPjRLPj5IHEHPMHHE^H1!Q@MtIIIMH8H5LIHVI%H$HML $LT$f.HHEt&H 3*H=y@Hx fDHF HHD$HFH$H~1HL VMH5l{Lwd*,fHFHH$PI7H5qLHV %HtHD$IGLm~ ff.AWAVAUATUHSHH(L5oLfdH%(HD$1HH$Lt$HD$HcIQI7IMHH HNAL HELNODHHnH ATH5H81'X0+ZH G9H=:E1HD$dH+%(H(L[]A\A]A^A_@LLMHHHsHH=[AHEHAVj55j5}Pj5IHEHPMHHEbH,UHV(HE nLIIKIMIHFHH$IMjL $HD$HT$IHF(HHD$HF HD$HFH$IM~1HL MH5vxLty+cHHEtFH Y+H=R@H AL HfDH(H5LIHVI!H$HLekHF HHD$HFH$IMH5LHV!HHD$IH5yLHVm!HtHD$IAWAVAUATUHSHH(L5!lLndH%(HD$1H$HD$Lt$H~I\IHV(HE LMHHHsHH=AHEHAVj55j5Pj5IHEHPMHHEHD$dH+%(H(L[]A\A]A^A_H`H5iLIHVI& H$HLmDIH YHJAHMEIHHjHH5E|AUL 'H81R#X+ZH H=E1p*LHhII[qM IHFHH$nIH5lLHV8HD$H3IML $HD$HT$,f.IHF(HHD$HF HD$HFH$IM~1HL ʀMH5tLqy+HHEt&H ~+H=I@HHfDH ~A`fDH5LHV5H`HD$IMfDHF HHD$HFH$IHHhH 9~H5NzjL 0AHH81N!Y^+mH }Aff.AWAVAUIATUSHH8L5ahLfdH%(HD$(1H$HD$HD$Lt$HEI+IIHj}H [}HNHATH~L@HgH5VyL :H81e XY,ZH }H=KE1HD$(dH+%("H8L[]A\A]A^A_DHV0HK(HC LKIHIuHH=bAHEAVj5UQHj5aPj5PbIHEHPMHHE]HPLtIHHJcH>fHF0HD$HC(HD$HC HHD$HCH$IItlI~.IML $HD$HL$HT$M*IH5vHHVBHD$HqIH59HHVHD$HIzfDHHEH ,{9,H=j~H5!HHVHtHD$IM#1HL |MH5 pHlE,fDHHqHFHH$pIHXIH5^HIHVH$HLcHHdH JzH5_vjL AAH{H81_Y^;,HHxdAH5vjL H yH8H{1_5,AX8AWAVAUATUHSHH(L5AdLndH%(HD$1H$Lt$HGI5IMH]yH NyHNHL ELNL@HHcAUHzH57uH81MX,ZH x>H=c|E1kHD$dH+%(H(L[]A\A]A^A_DLV LMHHHHsHHkH=AHEHAVjRPjRLPj5KIHEHPMHHE^HQ@MtIIIMHH5LIHVIH$HML $LT$f.HHEt&H w-H={@HfDHF HHD$HFH$7H~1HL yMH5llLTid,,fHFHH$I7H5LHVHtHD$IGLm~Lff.AWAVAUATUHSHH(L5QaLndH%(HD$1H$HD$Lt$H~I\IHV(HE LMHHHsHH=AHEHAVj55j5Pj5IHEHPMHHEHD$dH+%(H(L[]A\A]A^A_HH5iLIHVIVH$HLmDIH uHzuAHMEIHH_HwH5uqAUL WH81Xl-ZH %uH=xE1*LHII[qM IHFHH$ IH5LHVhHD$H3IML $HD$HT$,f.IHF(HHD$HF HD$HFH$$ IM~1HL uMH56iL>fyZ-HHEt&H t-H=wy@HxfDH sA`fDH5LHVeH`HD$IMfDHF HHD$HFH$O IHH]H isH5~ojL `}AHuH81~Y^P-H *sAff.AVAUATUSHHL-]HndH%(HD$1L,$H5H$HHHrH rHHHH|H?L HLHL@HH\UHKtH5nH81X-ZH LrH='vE1HD$dH+%(lHL[]A\A]A^HVHHHsHE1HۿL ̿HEH=`HAUjPAQjPAQjPIHEHPMt`HHEsHB fDLIHHHFHH$J HGH$P@HHEt&H Hq&.H=u@H 1HL rIH5eL cy-fH IHwH5LHVHtH$IFM2 fAWAVAUATUHSHH(L5A[LndH%(HD$1H$Lt$HGI5IMH]pH NpHNHL EzLNL@HHZAUHqH57lH81MXk.ZH o)H=sE1kHD$dH+%(H(L[]A\A]A^A_DLV LMHHHHsHHH=AHEHAVjRPjRLPj5KIHEHPMHHE^H Q@MtIIIMHH5LIHVIH$HML $LT$f.HHEt&H n.H=r@H fDHF HHD$HFH$7H~1HL 7pMH5 cLT`d[.,fHFHH$I7H5LHV HtHD$IGLm~L ff.AWAVAUATUHSHH(L5QXLndH%(HD$1H$HD$Lt$H~I\IHV(HE LMHHHs H=AHEHAVj55j5Pj5IHEHPMHHEHD$dH+%(H(L[]A\A]A^A_HH5YLIHVIV H$HLmDIH lHzlAHMEIHHVHHnH5uhAUL WvH81X.ZH %lH=`pE1*LHII[qM IHFHH$IH5\LHVh HD$H3IML $HD$HT$,f.IHF(HHD$HF HD$HFH$$IM~1HL ,mMH5_L>]y.HHEt&H k#/H=6oy@HxfDH jA`fDH5ɸLHVe H`HD$IMfDHF HHD$HFH$OIHHTH ijH5~fjL `tAH4lH81~ Y^.H *jAff.AWAVAUATUHSHH(L5TLndH%(HD$1H$Lt$HGI5IMHiH iHNH L sLNL@HHSAUHckH5eH81 X~/ZH @iH=mE1HD$dH+%(H(L[]A\A]A^A_DLV LMHHҶHHsHHH=AHEHAVjRPjRLPj5IHEHPMHHE^H!Q@MtIIIMH(H5ALIHVIH$HML $LT$f.HHEt&H gT /H=Vli@HhfDHF HHD$HFH$H~1HL iMH5\LYdn/,fHFHH$@I7H5aLHVHtHD$IGLm~ff.AWAVAUATUHSHH(L5QLndH%(HD$1H$Lt$HGI5IMHfH fHNHL pLNL@HHPAUHzhH5bH81 X0ZH PfY H=jE1HD$dH+%(H(L[]A\A]A^A_DLV LMHHHHsHH˳H=AHEHAVjRPjRLPj5IHEHPMHHE^H1Q@MtIIIMH8H5QLIHVIH$HML $LT$f.HHEt&H e +0H=iy@HxfDHF HHD$HFH$H~1HL fMH5 YLVd/,fHFHH$PI7H5qLHV HtHD$IGLm~ff.AWAVAUATUHSHH(L5NLndH%(HD$1H$Lt$HGI5IMHcH cHNH-L mLNL@HHNAUHeH5_H81X0ZH `c H=hE1HD$dH+%(H(L[]A\A]A^A_DLV LMHHHHsHH۰H=4AHEHAVjRPjRLPj5IHEHPMHHE^HAQ@MtIIIMHHH5aLIHVIH$HML $LT$f.HHEt&H b' 0H=f@HfDHF HHD$HFH$H~1HL cMH5ULSdv0,fHFHH$`I7H5LHVHtHD$IGLmff.@AVAUIATUSHH HL%KdH%(HD$1HHnH$HD$Ld$MHHHMLNHHHs H=ZAHEHATj5jRLj5.Pj5WIHEHPMHHEtIHD$dH+%(H L[]A\A]A^DHLV(HF UfHfDIM5DHF M HH _H_IHHI?IAHH&JHaH5[UL iH81X1ZH y_, H=\dE1H ]_AfH^HHIHFLH$IM~TH5LHVHtHD$IM~.H5LHVH HD$IML $HD$LT$H|H.HF(LHD$HF HD$HFH$DIHHEt&H H^ D1H=&c@HfDLIHSH5LHVHH$IHF LHD$HFH$I1HL _IH5QLO1fAVAUIATUSHH HCL% HdH%(HD$1H HnH$HD$Ld$MHHHMLNHHHs H=:AHEHATj5RLj5~Pj5MIHEHPMHHEtIHD$dH+%(H L[]A\A]A^DHLV(HF UfHfDIM5DHF M HH &\H'\IHHI?IAHHvFH^H5XUL eH81&X1ZH [ H=`E1DH [AfH^HHIHFLH$=IM~TH5LHVHtHD$IM~.H5@LHVH HD$IML $HD$LT$H̨H.HF(LHD$HF HD$HFH$IHHEt&H Z 1H=_@HfDL@IHSH5MLHVHH$IHF LHD$HFH$I1HL \IH5NL L1RfAVAUIATUSHH HL%\DdH%(HD$1H]HnH$HD$Ld$MHHHMLNHHHs H= AHEHATj5 RLj5ΦPj5IHEHPMHHEtIHD$dH+%(H L[]A\A]A^DHLV(HF UfH8fDIM5DHF M HH vXHwXIHHI?IAHHBHhZH5hTUL KbH81vXE2ZH X H=L]E1H WAfH^HHIHFLH$IM~TH5^LHVRHtHD$IM~.H5LHV,H HD$IML $HD$LT$HH.HF(LHD$HF HD$HFH$IHHEt&H VT n2H=\a@H`fDLIHSH5LHVQHH$IHF LHD$HFH$?I1HL tXIH51JLYH12fAVAUIATUSHH HL%@dH%(HD$1HHnH$HD$Ld$MHHHMLNHHHsHH=AHEHATj5ZRLj5ާPj5ͧGIHEHPMHHEtIHD$dH+%(H L[]A\A]A^DHLV(HF UfHfDIM5DHF M HH THTIHHI?IAHH?HVH5PUL ^H81X2ZH iTY H=YE1H MTAfH^HHIHFLH$IM~TH5nLHVHtHD$IM~.H5LHV|H HD$IML $HD$LT$HlH.HF(LHD$HF HD$HFH$4IHHEt&H 8S 3H=X@HfDLIHSH5uLHVHH$IHF LHD$HFH$I1HL TIH5aFLD2fAVAUATUSHH L L-<dH%(HD$1HnL $Ll$HQH?H%HHHRIH RH-THIHHb<I?UIH5NH8L [A1Xb3ZH Q H=DWE1,HD$dH+%(H L[]A\A]A^MHHGHHs H0H=AHEHAUjRPjRLPj5ޞIHEHPM|HHEbHULV LNvMIHHteHHIHL $LT$%fHHEH |P 3H= VHF HHD$HFH$'H~1HL rRIH5CLDByQ3_fDHFHH$IMTH5LHVHtHD$IF느HhAH5yLHVmHtH$IfAWAVAUATUHSHH(L5!:LndH%(HD$1H$HD$Lt$H~I\IHV(HE LMHHHsHH=ȨAHEHAVj5؜5ʜj5Pj5IÜIHEHPMHHEHD$dH+%(H(L[]A\A]A^A_H`H5LIHVI&H$HLmDIH YNHJNAHMEIHH8HaPH5EJAUL 'XH81RX3ZH M H=SE1p*LHhII[qM IHFHH$nIH5DLHV8HD$H3IML $HD$HT$,f.IHF(HHD$HF HD$HFH$IM~1HL EOMH5@L?y3HHEt&H L_ 4H=RI@HHfDH LA`fDH5LHV5H`HD$IMfDHF HHD$HFH$IHH6H 9LH5NHjL 0VAHMNH81NY^3mH KAff.AWAVAUATUSHXL-L5xH|$HndH%(HD$H1H@6Ll$0Lt$8HD$@HHFHH! H5HD$LnH H(hE111HALIH[ H8qHz H(hE111HALHH H8]EA;D$/H@H H9H LM IIBLT$LH5HH LT$HIHH IHH{E1E1AH;=[4] H;=4hH595H9tLD$LD$tHSBፁ LLD$LD$HI; MtL@IcHEAIlHCMcI$OdLM( H=D1LHAI9MXI/H+lHIL0IH| H8RHHH9X LۜM IIBLT$LH5HH LT$IM I*H7HxH9X H_H HHCH5HHHT IMn H+Hu2I9B LLLT$c3LT$HH:I*H92I9GwHL,3IH+IHMIHL;t2L;B2u L;T2nI*! YHL$HeMHAH=LHq IL52jP5(jPAUjPHT$XLT$` HPLT$HI I* I,$MHmMt Im Mt I.HD$HdH+%(,HXL[]A\A]A^A_HfHF(Lv HD$K@H(Lf.GD$zLf.Gz\D$H͕H>D$ H9XLM~ IIBLT$(LH5GHH LT$(IIHMU IHAD$ 6IH H/I9G LLLD$(0LD$(II(7M I/L;-0L;-/u L;-/?Im Lt$D$MIHHD$ wIH;H̒LILH5/AH H=pjP5jPAUjPHT$XIHPH I/:H+ImE1E1Yf.HEA;D$DH.HD$H.Lv HD$L*HE1LLD$Ld$HDL`LPqLHLD$0L)LD$Ht0Hl$8Ld$@+LD$ &HIMI( LfDH ECAHH-HkEH5J?UL -MH81XX%ZH BH=HE1vIHHTHHFHHD$0IM~TH5LHVNHtHD$8IM~.H5LHV(H HD$@IM HD$@Ll$0Lt$8HD$HH 6BH7BIHHI?IA@HHF(HHD$@HF HD$8HFHD$0IfDHhHXE1E11E1E1ҺT<%E1MtI*tJMtI(tgMtI/t|H nAH=^GM}I,$]E1jDLLD$T$t$LD$T$t$@LljT$t$T$t$|LT$t$T$t$gE1E1E1E1ҺUK%%DLPHLT$;LT$ML(=LL[HLT$LT$jfLLLT$LT$yit&E1jfLxÅMﾝ%[E11E1E1fDI/2 LT$t$LT$QL|$IE1t$T$fH(E1E1Xi%E1E1D$ D$ HE1E1Ys%E1E1H=H"H5#ΥIME1E1E1[%Rf.LLT$sLT$fLXOH=9HH5VIME1E1E1d%fLL"d%ME1E1E1I(tE1E1fLljT$E1t$LT$t$T$E1LT$^H=y$II@LT$HpfDLCMLkIIEH+I}LAAg@&MIE1E1d%1LH IHE1E1$&dSLHD$LD$fLD$01L)Hl$8Ld$@Ht0LR uH{LD$LALD$Iƾ&MG-E1E1Һh<&LLD$E1E1LD$ZHLD$LD$LHD$LD$fH=H H5 ֢IM$E1iK&`LT$I%fDH=a IMкiE1ҾM&H=1HrH5sNHHE1ҾP&iHHIHH5ňLHV HHD$0IfHF HHD$8HFHD$0IfDIH=,HY@IؾR&iI/LLD$ T$t$LT$OLT$t$T$LD$ ^H=ĠIq@[%#fKLT$(IqfDMBMIJIHI*LLHLD$ HL$M#LD$ LT$HI(LLT$LT$a&if$&dE1E1E1E1E1M[%MGM|IOIHI/0LHHLD$ HL$"LD$ L|$II(ELLT$LT$.1۾q&i/CIHH#H5I4H8E1Ҿ$&dMwMI_IHI/LLHLD$(!I.LD$(IILLD$(LD$(fDE1E1M[%L=iH5jIGHHXH=W3Ht$Ht$1LHD$LT$MLLT$LT$I*j&E1E1L-يL5ڊIEHHH=2"1LLHD$6LT$MbLLT${LT$I*TE1E1\%E1E1yk&iE1E1E1%_HLD$LAjI}LD$AfDE1E1%`MIE1E1%^1HL$0ILL 8H5,( $MLHL$ LD$LD$HL$ LHL$ LD$LD$HL$ LLD$(LD$(`LjE1E1Ҿ&@LpL1ALD$I1LlIHE1j&1LLAIH-E1E1E1\%j&E1E1E1E1\%E1E1xLT$HzHD$H>H5w0H8LT$LD$j&N1LT$H^HD$HE1E1H5*0H8LT$LD$\%ME1IE1E1E1IںdAWAVAUATUHSHhH^H|$dH%(HD$X1HHD$0HD$8HD$@HD$HHH HN HF0HD$H](Le HmHb L(hE111HAHAIH6H8 H L(hE111HALAIH H8O H؀ L(hE111HAHAIHDH8 AFA;E HH H9HH-HHEHEH5 HHHIMHmH9H :H9HH-!HHEHEH5HHHIHEHMHEHJI{1E1H;=bH;=BH5;H9LT$ L\$L\$LT$ HLT$ L\$L\$LT$ HIMtLPHcIEMlICHcIMtHHpH=,LL$ L\$ L\$LL$ 1LLH&HL\$LL$ I)I+ HPI9D$MHLBIHm MI,$ L;L;_ L;m LLT$*LT$Hc؅ I* xHH H9HL%MxI$ID$H5LHHDHHI,$ HH H9HsL%|MI$ID$H5LHHnII$HMCI$H IyE1H;=AH;=S H5H9tLL$ L\$rL\$LL$ tIAPуፁ+LLL$ L\$ L\$LL$ HI[MtLXHcIMtIAHcIM|HHxH=5*LT$ LL$vLL$LT$ 1LLIMLL$LT$ I*I) HH9ELHII,$ MHmL;L; L; LLT$LT$I*.H}H XH9HqH-?HHEHEH5HHHIM{HmH#}H H9H H-ˀH~HEHEH5HHHNIM,Hm^IzE1H;=XH;=H51H9tLT$ LL$LL$LT$ tIBPуፁHLT$ LL$LL$LT$ HIMtLHHcIEMlIBHcIM|HHH='LT$ L\$L\$LT$ dL\$ 1LLLT$HHLT$L\$ II+I*pH,I9D$HLIHm.MI,$L;mL;;` L;IS LLT$LT$I*uHD$HML :yH= LHp ILjAQAWjAQAVjHT$PLT$H;}H@LT$HHI*LLIH5>}LIHVHD$0HH]@HH*H *HNHSH,L@HH5&L 4H81X4ZH X*d H=k0E1ӝHD$XdH+%(HhL[]A\A]A^A_DHHD$HKHL\$L\$fHLLT$0H)LT$ 5Ht0L\$Ll$8Lt$@L\$LT$ HHMRI*HLL\$L\$1f.ISBፁWLT$01H)Ll$8Lt$@Ht0LB uI{LT$ L\$HALT$ L\$HŻ5HWMI*1E1ɺ I+Mt I,$Ht HmMt I)H {(H=.Mu7Mt I.(M I/LImE1HoHA;GHVf.V)f(zHL$ 4f.4)L$ D$zLL$ f. )L$ f(zf/f/T$ f.L$z H\$f(T$ LIL\$L\$HIw T$ HD$(f(L\$LL$(HI= D$LL$ mL\$LL$ HH HtH5vLjH APHH=lUjPATjPHT$`LL$hL\$X~tHPL\$LL$HHD I+I)I,$eHmJImt5I.IfDLH1LIfDVL߉T$LL$T$LL$ LT$LL$_T$LL$HT$LL$?LL$T$LωT$$T$HIHH,HcH>f.HF0HD$HHE(HD$@HE LHD$8HEHD$0IHtsH~5H|MRHD$HHl$0Ld$8H\$@HD$@HH>H5&wLHVzHD$8H0IH5vLHVUHD$@H IsfDE1E1E1ɺ 4E1E1ɺ 4HLT$۾LT$=LȾ E11L׉T$LL$L\$蠾L\$LL$T$M.3@E1ɺ 4>fDLLT$[LT$H5qLHV]HtHD$HIM1HL$0ILL %H5\w4@MSM$I[IHI+ H{I۽HFLHD$0߻I 6E1MJI*@L׉T$LL$_T$LL$!LHL(`LL\$L\$@fH=qHuH5uIMtLWfDE1ɺ 5fDE1ɺ 5efDH=q4IILLL$kLL$ H=IqHJuH5KufHHE1ɺ 5LLL\$0H)L\$ D6Ht0LL$Lt$8L|$@z LL$L\$ HIpMI+LLL$贻LL$uf.H蘻L舻LxsLLL$cLL$OfLLT$CLT$8fL(H= p贆H@SIIE11 5I*E1@HLT$úLT$f1LLLL$ L\$L\$LL$ HH 5@LhI\$HIL$HHI,$aHHHHL$H+Ld$IvHLT$LT$_fDE11 6@D$uL$HD 5E1t@L$ EL$ H> 5E1D@L$ D$(L$ T$(H* 5E1H-mL%pHEHH H= 1LHI(M LwI,$B .5E1f.H-9mL%RpHEHH^ H=7肷 1LHI訽M LI,$! 6E1 f.H(H-lL%oHEHH9 H= 1LHIM LgI,$ N5E1zf.LLL$蓷LL$TfH=ilHZpH5[p膃IM_E1ɺ +65fDH-kL%nHEHH H= 1LHI(M* LwI,$ n5E1f.E1ɺ -6MfDHH=qkI@HLT$cLT$f1E1亿 5@H=!kHoH5o>IM|E1ɺ 06 51HcӸLLL$0H)HLL$ 6Ht0LT$Ll$8L|$@?LT$LL$ HHMI)LLT$tLT$f.1 5I)E1HLT$5LT$L#M E126TIH=i萀ILL$5L\$脹L\$LL$ HaHEH5~T$H8L\$LL$T$4 5,LLT$sLT$ R6 ME1MYMIYIHI)=H{IA1LLLT$ LL$LL$LT$ HI ]6 5mL\$01H)Lt$8L|$@Ht0H@ uIyLL$ L\$# LLL$ L\$IĻL6MLVDME1E1ɺ o6H]H HMHHHm(LHHHL$H+Hl$IHLT$޲LT$ r6E1?LLT$谲H{ILT$H-#gL%4jHEHHpH=!l71LHI蒷M.LI,$ 6E1LLT$LT$hH=fHjH5j~IMuYE1ɺ 6LHL$̱HL$E1ɺ 6oI]H=f7}IL!L.5} E1@HHEH H5jL AHH81Y^m4-L6 E1H=eHiH5i}HHE1ɺ 6 ]6E1LN5躰 E1}I 16HIH=ue |HLT$]6LL$LL$LT$ HiHH5T$H8腱LL$LT$T$ 艧LLLAI譭MtI.I,$Hm]LA\A]A^A_HIMHI9Fu|M~MtsMfII$I.LLI/IthI$HMI$HkL貧HmcH蟧]LA\A]A^A_fLMuILxLhfDLXgHI腥H1H5jL+E1AA&DDE1H ] H=A)'I.AI,$uLæ뵐LLLҮIHBfDA*'ۯIaAA&cAA%'K蛯I_AA ']I.u[LAA''fL!蛪HBHkH5H8'AA''fDAWAVAUATIUHHHEHEHIT$H5]LHHHIMsƫIHH5XHHجIFL=EYLMH=LLLAIMtI.I,$Hm]LA\A]A^A_HIMHI9Fu|M~MtsMfII$I.LLI/IthI$HMI$HkL"HmcH]LA\A]A^A_fLMILLأfDLȣgHIH1H5` L{E1AAw'DDE1H H==P|A'I.AI,$uL3뵐LLLBIHBfDA'KIaAA'cAA'K I_AA']I.u[LAA'~fLh! HBHH5H8茣'AA'fDAWAVAUATUHSHHL-H^H<$dH%(HD$81Ll$(HD$ Ll$0H( HnHLHHHH HNAL Ht-ELNODHHSH/ SH5H81 X(ZH H=@ E1(zHD$8dH+%(5HHL[]A\A]A^A_fLl$MLuIHUI$M9H ZH9H H-YH HEHEH5THHH HHEHHQ HEHH=XH5THGHHj HHl ϦIHs H5 SLHAH5RLLǧ/ HEH54TLMD H=Ht${Ht$LHAHD$LT$MUHmI/H{E1E1H;="<H;= H5H9LT$轧LT$HLT$sLT$HHMtLxIcALTHWMcHJDHCLMH=61HHAIM0Hm H+ImH<$H5PHGHHg IMq H-I9l$LHH H@H5OH9 HCHPH} H\SH)HJHHEHHHl H+rHHT IHh LpKHH HT$H5+QH[K HVH5}PH=U IELM H=e HHLAI艢MIm H+ Hm I.FI,$LfDH !VH9H HVHWHHCH5PHHH HHHHP HHH= UH5PHGHH HH\ ߢIHc H5OLHHD$LT$HCH5TPLM H=LT$Ht$LT$ LHt$HAIMLT$ H+I*H}1E1H;=Af H;=qH5H9T$L\$ӣL\$HcT$HL\$T$耟L\$HH6 MۋT$tLXHcƒL|HSHcHHDHELM H=1HHAHD$LT$M, H+ Hm<I*"IImH- LHEI.MIf.HF(HD$Le VfDLl$HSBፁH RL|$ 1LT$(HL$0LZL)Ht uH{LT$ HALT$I15A(MMt!I/uLLT$辙LT$fI*L螙fHEH΃捁H5QL\$ 1L|$(Ht$0H@H)Ht uH}L\$5 HL\$ILۺ.A(MMt I+ I/fDHLT$LT$2E1/AZ( DA)){DfDH5yBLHVŔHHD$(IHHL蠗IH4A-)f5A(/A](E1mDL]MH]IHHmH{Hݺ`1HHIHE1.A(L$yL$/A](Hm5A )E1E1AHH$H H5FE1H8軏L$5A )A-)HHH5D$H8yD$LLT$LT$Lۺ.A({CHH HCSHH H)HLT$HA谍H}LT$L$DL$HIH H5FH8辎L$.A(HL\$HJH{L\$ XCH1HALT$I1HL\$Iʍ:E1E1 f.AVAUATIUHSHH HLndH%(HD$1H$HD$HT$M}IIIH HAHMEIHHHH5AUL H81˔X;ZH ng H=E1dHD$dH+%(H L[]A\A]A^DHV(HM HELHHsHE1H=+{AI$j5>5>j5ZFQLj5FFPFHI$HPHCHI$t)HGIHt)HmEH68L(fDH z ;H=c븐IlMSIHFLH$IH5<LHVۏHD$HnIMH$HL$HT$DIHF(LHD$HF HD$HFH$蜈IM~1HL MH5Lyp;HI$t&H p ;E1H=bLfDH EAfDH59=LHVՎHXHD$IEfDLЇH5iDLIHVI薎H$HLmHF LHD$HFH$臇IHHH H5jL AHhH81趑Y^f;ՉH bAAUATIUSHH(HHndH%(HD$H;HT$H$MHHHHHIH HHIHH3I?UIH5H8L A1ߐX<ZH  H=-E1`HD$dH+%(H(L[]A\A]HV HFHHLS E1H5;H ;AHEH=^wjVQjVLQHj 50BP9BIHEHPMHHEt+LCIHt+I,$KLP>H@fDH  O<H=8 `fHHtpHzL/IHH$HT$fHHEt&H A<E1H=_H蘆fDHF LHD$HFH$跄H~1HL IH5Ll<fHFLH$pIMDH59LHV1HtHD$IE느H5@LHV HtH$I蹆fAVAUATIUHSHH HLndH%(HD$1H$HT$MIIMHH HNH<L LNL@HHAUHH5H81̍X<ZH oH=BE1]HD$dH+%(2H L[]A\A]A^fDHV H}LHHs E1H 7H7AI$jQPjQLPj57WH=B$?HHPI$HHI$tNH@IHtHmIH;-LIHVwHD$0HcH]@HHH HNHSHL@H5H5L H81zX=ZH H=E1 KHD$XdH+%(HhL[]A\A]A^A_DHHD$LqRLqIAH΃捁QIcվH\$01H)Ld$8Hl$@Ht0H@ uIyLL$LLL$IE1ۺ?MHH+HLL$!qLL$xIFPуፁH $LL$01Ll$8HL$@H@H)Ht0 uI~LL$DLLL$HHMt I)bImLpfDHppH`p*H $;AnfIF2 MFI@Hk HuHcH>EF@IG MOIAH8 HFHcH>@EO@HC LkIEH: HHcH>@Dk@KL9 HD$LLL$LIrHHH|$rIHDLlrIHHt$HE1LjH=TA5l)H PHD$(j5R)AWj5@)SHT$`|)LT$XHPHII.U H++ I/ I*L+IH:I,$tOHmMt^fH $HHD$HHtsM7Im,L?nf.E1LM"nfHtHmu H n@H<$uHmHmfDHFIH\HcH>f.HF0HD$HHE(HD$@HE LHD$8HEHD$0kIHtsH~5HMbHD$HLt$0L|$8H\$@HD$@HHNH5'LHV:rHD$8HrIH5'LHVrHD$@HIsfDH$E11E1E1E1>E1E11Mt I/Mt I**Mt I.SHtH+tpMt I)MtI+t6H H=3EMI,$E1L߉T$t$lT$t$fDHL\$ T$LL$t$kL\$ T$LL$t$_LL\$T$t$kL\$T$t$@LL\$(T$ LL$t$LT$kL\$(T$ LL$t$LT$DLL\$ T$LL$t$FkL\$ T$LL$t$LL\$ T$LL$t$kL\$ T$LL$t$|E1E1E1E1H$->DHLL$jLL$E1E1E1E1 <>DIcոLLL$H)LH\$0Ht0Ld$8Hl$@E1LL$HIž ?E1I.tE1E1@LL\$(E1E1T$ LL$t$LT$jLT$t$LL$T$ L\$(?LoÅE1f?E1E1fDEFAMcILD$BnLD$HE1 ^>E1E1E1fEFAFII I@EFAFII 4@EOAMcI\LL$ LD$mLD$LL$ H:E1h>E1E1E1/EOAGII I@EOAGII @DSAMcILL$ LD$5mLD$LL$ HE1r>E1E1E1DSCII IMf.DSCII Mf.LhfH5qLHV mHtHD$HIMt1HL$0ILL CH56L=n@LgH=yH#H5#3HHE1E1ɺ>DHFLHD$0eIL(gLg~H$LLLL$0LL$HD$@H)Ll$8Ht0pLL$6?HH%ImAt)E11qfDHE1E1Ҿ>LT$ LL$t$LT$ufMt$T$ LT$LL$uE1E11HHfL8fL(fLLL貶If.LeH=H!H5!1IME1E1E1>fLe LLT$eLT$fHLT$seLT$fLXeLT$fDME1E1>H=0II@cnI~LLL$dLL$RH=H H5 0IM~E1E1Һ>L}iILLD$hiLD$IeH=q0ImI:HLL$ LD$!iLD$LL$ IHLLT$ϴLT$MI,LLT$ dLT$HLT$cLT$fLLT$cLT$f mIH=D/H)@E1ۺ>FfDIYHMiHIEI)I}MAA1LLL\$ LL$MkLL$L\$ HIn1ۺ"?fDLDIJLLD$DLD$IfDHLL$ LD$DLL$ LD$I@L51L=IFHH H=/za 1LLIgM LImE1>E1E1E1yfMNMI^IHI.H{IAmHLLT$vLT$MH1LLL\$iL\$HHE1ɺQ?QLLT$aLT$!fHLT$ LL$^aLT$ LL$fE1E1E1>LLT$#aLT$CfHLT$aLT$;fE1E1Ҿ>H+I?fDE1E1Һc?MoM!IOIEHI/HLHHL$ImL|$IL^`fLLT$C`LT$fE1ɾ>>fDLLT$ LL$`LL$LT$ ?E1ɾ>fDL5QL=2IFHHH=O^k1LLIdMkLIm^E1u?E1E1E1E1E1Һ>L,_LLT$_LT$HLT$ LL$^LT$ LL$FE1E1E1Һ?'"hIF?"?1oLLT$^LT$HLT$^LT$H=jHH5*IME1E1Һ?L\$LL$bLL$L\$HH HH5H8[_LL$L\$"?LLT$ LL$]LT$ LL$LLT$ LL$]LL$LT$ H=D)I8ME1?fILr]L$hQ?E11E1E1Iߺ?L\$aL\$HHD$HH5L\$H8U^LL$L\$Q?L-%M IrI9HXH$HJH~1@L;lHH9uHVIMLT$E1HeH5E1H81!eLt$E1Ҿ?f.ME1E1E1Һ?eLT$IJH= HLT$H5$(LT$IMPME1E1Һ?+L[rLLT$ LL$[LT$ LL$H=LT$<'LT$IME1E1M?7dLT$HL][HH-H H5ϼjL AHH81cY^=LLL$I[H{LL$AqME1E1MϺ?L- MSIvI9HXHkHJH~1L;lHH9uHVIMLT$E1HEH5E1H81cLd$E1Ҿ?f.H<$wcLT$ILME1E1E1Һ?LLHL$YHL$YH=H?LT$H5;%LT$IM[LME1E1Һ?HHoAH5jL H H8H1b_=AXLME1ME1?bLT$HH= LT$$LT$IHLX1LL aIHE1E1E1>LME1E1E1MϺ?LMMH$E1E1 ?L- MhHsI9HXHHJH~1DL;lHH9uHVIMLT$LHIE1E1H5H81`Ld$E1Ҿ?<1LL_IH\E1E1E1q? 1LLL$IPE1>E1E1E1E1L??1LLL$H\IHrHۦH5E1H8XE1Һ>wE1q?E1E1E1\[IH(HH5E1H82XE1Һq? HYH5ϼLT$E1H8XLt$E1Һ?HHI9wHuL;-eLWHH5kLT$E1H8WLd$E1Һ?HHI9[HuL;-XIHH5 LE1LT$IH89WLd$E1Һ?"HHI96HuL;-$tE1E1E11E1>ff.AWAVAUATUHSHLfH|$(dH%(H$1HUHD$pHD$xH$HdHD$8H$HHHD$@H$HIIIHE(HD$HE L}HD$HD$IHRIHHIQLLL$HHH=1u"LL$HH$ I)jH5kH[#IH/ H8}HmbH H|H9XR L%cMI$ID$H5LHHHHI,$mHVH9CLHIIM=H+kI/QHzHH9PL%MyI$ID$H5HLHHHHI,$ HH9CHt$H覣HD$H|$UH+H\$HHD$ HHHD$H;բHD$H@ PIH H\$H@Ld$HHIFH5cLHHhHHjHWIHHmIhH|$H5 HGHHHHHrWHD$HH;Hm H|$H4H-UL% HEHH-H=SP1LHIVM3LI,$$HD$AMBHD$HD$01E11HD$(E1E1HD$ HD$]DI^HF8HD$@HE0HD$8!LPHHD$HPHPHm{H H9HXHt,HqH~K1HH9t7H;TuHH9HuH;|fDHD$HfH= IM HD$0M11HD$(E1E1AiAHD$ HD$HD$Mt I+IMt I*Ht H)H DH=zm(Mt1Im#IHL$HtHHD$8HHMt I,$Ht H+H\$HtHHD$HH HT$ HtHHD$HHHL$(HtHHD$HHHt$0HtHHD$HHI.H|$HHD$HHH$dH+%(8HĨL[]A\A]A^A_fDLPNLI=NH(NLNHNHMHMHMHMI.LM H|$ML׉T$@HL$8MT$@HL$8HωT$8tMT$8L߉T$HHL$@LT$8RMT$HHL$@LT$8L0MH MLMHMHLIIHJcH>f.HF8H$HE0H$HE(H$HE LHD$xHEHD$pJIIH2JcH>@LJIH56LIHV^QHD$pHlLeIH HAHNEOD@HHޛHH5ATL bH81TX AZH 0H=E1$,fDH A또M1E1E1HD$01E1E1HD$(ASAHD$ HD$HD$HHKHD$0M11HD$(E1E1E1HD$ E1E1AXAHD$HD$MtI)tWHt HmMI(LL\$PT$HHL$@LT$8JL\$PT$HHL$@LT$8LL\$XLD$PT$HHL$@LT$8pJL\$XLD$PT$HHL$@LT$8lHL\$XLD$PT$HHL$@LT$80JL\$XLD$PT$HHL$@LT$8<M1E1E1HD$0E11E1HD$(A[AHD$ HD$HD$H=H H5 IfDH|$H5|HGHHHHHEH;H;H@hHxH@Hk1HIM$HmH|$LL$H5HGHHLL$HHXHEH;qH;tH@hHAH@H4LL$HLL$IM5HmLLϺLD$ LL$ILL$LD$ HHyI)/I(HH;-?H9H;-HMAąHm;EIFH5LHHHHHEH;KH;NH@hHH@H1HIMKHmH|$LD$H5?HGHHLD$HHHEH;H;RH@hHgH@HZLD$1HLD$IMpHmHLLǺLL$ LD$^GLD$LL$ HHI(I)H;-H9 H;-u H7LD$P3HmL$PH|$HWHBpHH@HH5HH&HMIHoHmmIFH5LHHIMIAH;FpH;IH@hHGH@H:LL$L1LL$HHI)2IT$ ID$HHH9;H92IT$HHHID$H+H|$(H5HGHHkHHHsH9CCHkH6LCHEIH+LLHLD$:HmLD$IjMI(IALL$LH5AHHLL$IMI)IFLD$LH5HHLD$IMIAH;QH;H@hH H@Hs LD$(L1LL$LL$LD$(HD$ H|$ f!I) IxH;=!MHM-"IhIHEI(_H}AH;=eH5ȓH9tLT$(LL$KLL$LT$(tHEPуፁ !LLL$!GLL$HI$MtLHHHcHIDCHT$ HITHEHHH=JLT$ALT$&1LHHGHLT$2'I*UHm$H|$H5"HGHH?!IM HHRH9P!H9HD$ Hs"HH|$ LD$H5HGHH1"LD$HD$(H|$(!HT$ HHD$HHyH)I9@MHMIhIHEI(HT$(LHLL$ LL$ HD$I)HT$(HHD$ HHH|$)"HmAHT$HHD$ HHIEH{I9E#Ht$LlMHHd#I(WHEH;X#H}H%HEHD$HE HD$ HE(HD$(HD$HHD$ HHD$(HHmH5H|$8m$HD$0HHH9PM(L M (IIALL$8LH5HH'LL$8IM_'I)HHH9P(L M`(IIALD$@LLL$8H5HH(LL$8LD$@IM(I) HՎI9C)Ht$ LLD$@L\$8輏L\$8LD$@HH)I+!HEH5iH@pHv%H@Hi%LD$8HLD$8IM)Hm>#Ht$(LLD$@LL$8 ?LL$8LD$@HH*I)#IxE1H;= *H;=$H5H9tLD$@LL$8FLL$8LD$@tI@@= *H|$HLD$@LL$87BLL$8LD$@HI*MtLHHcD$PHI\@HIlI@HH(*H=eLD$@LT$8HHEL@IHLD$ LL$4LD$ LL$Hh4SH9HRH5ABH81Ll$hHcT$1HLBIH1A= A\8H 1YRHH=_AD A8HD$pH|$hHJ[fHAR A9E11BH~AS A9E11AT A9E11HKLL$x>HH HxH90HXH HqH~1fH;THH9uHAHJHAD HWH5]VA8H81m8AE A8HQAE A8AU A9E11HAAF A8H8HAA8H8vHD$pHmAF H=HuH@H5QH8:HmAF A\9SHH=HH5ԼHH\$pHE1A\ A^:A= AQ8Lv1HHH+HuHSHr1A` A;L/DA9H=賻HZ1A\ A`:CIID$HD$pHIl$HAD$ HEHl$hI,$LHl$hzH?LH$A8L$H$H)HĀ<HD$hHt.Ht Hm]I,$L0A8I,$A@ tL gA= A\8E11q1HHHD$xHIA:1A\ kHHH8>H5qOA= A\8H8Lx#LkH >Hct$ 1L$Ld$pH$H)L$LBH̀ uH}D LAHH\$xHE1A\ Az:lHM>HA` A:H8IHxH@H|$HD$0:H>HA:H8HD$xH+lA` 1uH|$pHl$xIHCJ@HD$hIOA:H7UHH|$pE1A\ A:L1A= AC8PLLbL%MHHI9HXHAHrH~1L;dfHH9uHc<MD$1H5QHQA? A{8E1H8L1IA:A\ 1HD$xI,$LE1 HA` 1 HfIA:1A\ HD$x@H=HH5޷IMA@ A8!H+XA` A;1H|yHHL;H PH5LjL ZAHWH81Y^W7}IA@ A8A8A@ NI#H=ߟ芶I'1LLHD$hIHFA@ A8CInH5MnHEIELl$pI.LL$ HT$kLl$pHcL$ HT$HH+HA` A;1)1HH7H+HuHHE11AX A=:gHE1A\ Ar:FLH59H$1L$H$H@H)HA uI}AHD$hA8H&H9LE1AX A9H8jA9H+t@AX E11Hj9HA9H81HD$pHHAX E11AlHHD$hA@ A8)fA@ A8E1HD$hH MY 9H=VFH|$8HL$xHT$pHt$h HD$hHL$xHT$pHD$PH1HL$@HT$H(HHHCHHH=HHL$XHL$X1HHHD$XLT$XMZH+'HmML;7@L;h7@L;u7LLT$X2LT$XI*<HD$PHtH诔HD$hHD$HHtH蔔HD$pHD$@HtHyHD$8H\$HD$xHH8L`HXH\$0HhL(HXHt H/Mt I,$HHmHy@H5H5oLHAD A8H81HAI|H|$pHHH9%HuH;M6IHH+HAX E11A=:II*@}LhHLT$XLT$XH5H5TFH8HLT$XjLT$XA9HD$8HL$HH8HhHHHL$0HXL(HHHt H/(Ht Hm"HtpH+1AX sH+AX A=:E11S1LAHHL4H5J1A? A{8H8E1AX H4H5NEH8(1HHI9(HuL;%4AF A8E1H3H5DH8_H+ A:0HL$@HT$HA:Ht$PHSHD$hHD$pHD$xYA :NLA` A:HAX 1pH HAA:HtE11HHOIHD$XH2H5CH8\LT$XZA:HE1AX A9LA8A9H?2H5xCH8f.ATIUHSHHHt HՅuH1Ht[LH]A\[]A\ff.HGHHu710Ht!HSHPH2HHPHH@H!2H5 18USHHH-1HHEHkHHEHt H/tMHEHHHHEHtH/tH1[]cH1[]f.KfUHSHHHHt/HH/t H]H1[] H]H1[]fDH1ff.@AWAVAUATUHSHH0LnH|$ dH%(H$1H\$xHD$pH^ I$ Ij Lv HmHD$XHHD$`HD$hH$H H L(hE1ɹAHƺHAHD$HD$XHH8w HHHD$XHD$`H9PH-HHEHl$XHEH58HHHkHD$HD$HD$hH#Hm HYHHD$XH9PAH-`HHEHEH5?HHHyIMxHm I}E1E1H;=}.H;=. H5[/H9t !tIUBፁ5HHHzMtL`HT$IcAMcHHTH)HJDIELMH=>@1HLAIeMLd$XHm Im! HD$H5-H9p Ht$XH|$h|.HD$`HH|$XH/ HD$XHH|$hH/ HD$hH;--H;-z-H9HND$Hm HD$`L$cHD$LhI9 HD$0HDHL8Mt I9HPHuHXH@H\$HD$(MHD$HtHHD$(HtHHHH9XH-HHEHl$`HEH5'HHHILd$XMHm HD$`H+I9D$EH|$XL,HD$hHHD$`H|$XHH/ H<$H^DHY@I9uHt$0Hl$HD$HIH9D$(rLd$@Hl$HH|$8ID$L5HHH={71LLIMI,$qIm H\$HD$HHH$HHH\$Ht$HHt$H$HHH|$6II+ IMHMHz;H k;HNHaL bELNL@HH%AUHDH5T7H81jXHHZH ;H=@H苮HD$H$dH+%(pHD$HĘ[]A\A]A^A_@IHXH@H\$HD$(IUDIGPуፁ Hct$Lt$X1Hl$xH)Lt$pH$HtpH@ uILHD$HD$HD$hHMt I.pHD$XH+/H"‰D$@Hp|H`HPPHD$HLLd$pHD$xHH$L)Htp!HD$XHHS MI,$LLfDfDHLHD$1E11AH!E1HD$fH|$XHt H/tH|$`Ht H/xH|$hHt H/Mt I,$Mt ImHt H+H _8DH=E߫HL$HtHH$HHHD$HL$HtNHH$HHjHt HmJHL$HH$HHHy@HHmHT$@$L$%$DL$ $DH߉$$D$$|$$x1H|$H\$f.H<$HD$`HH HD$hHH HXHD$`HD$hHH1E1E11AH"HD$DHH=H*H5+HHl$XH21E1E1AH%HH5LIHVIHD$pH MHl$pLt$x+Hl HD$h1E1AI1@1E1E11AH%H=ʛH7mHD$L 5H=HH5 HH1E1E1AHHD$%OH|$`"HD$hIHEII11AHHD$%H=WHqLRH>HcT$Lt$XLHl$xH)LLt$pHtpH$HD$HD$hHBE1E1AI1vfDMeMImI$HEImH}IA1HLHD$XIHnHAH1%EHL$Ld$p1HL$xH H$LJL)Htp uI} HAHHl$XH1AH%L1E1E1AH%L`HMHxI$HHH|$hHD$HHAHT$XL=HD$`HI,$LHD$1E11AH%E1I$HF HHD$xHFHD$pvH1HL$pMLL ;H5}$v8HHHFHHD$p'IGH-hL%HEHHH=f-A 1HLHH Hl$`H!Hm.HD$`AH&1H|$X1E1E1HD$HL4P*3HH|$`HD$1AH%LD$`H5^LHVHHD$xIFLMH=HH5輗ILd$XM}1E1AI1KHAHE11HD$X%*H=>IAI1I=H=|HH5-HHl$`HjH|$XBIHt H/ HD$XH|$`Ht H/HD$`H|$hHt H/H 0-HD$hH=+=vH|$0HL$hHT$XHt$`=HI9FIH<$HHaIH:HXHLHHjI.I,$H|$`Ht H/HD$`H|$XHt H/HD$XH|$hHt H/HD$0H\$HD$hHH8LpHXH\$(L`L8HXHt H/oMt I.jM I,$L8HHAH1%HD$XLfLH=}HH5ILd$XM%1E1AI1{I*H=:z-H{DIHH=W}IHD$E1E1AI11LLQI,$IuLM~1E1AJ7SIIl$Hl$`HI|$HEHH|$XI,$t4fLHUHD$hHHm|HoLHl$`HDH|$XLpkLci1LLqHD$HD$hHfIFHD$XHM~HAD$IL|$`I.LL|$`VIL1AH%qH|$XXIH1E1A*JH87HaH|$hHA,JH8HD$`I,$71E1E1HuHH5&H8LI,$uLAJA:JAI.1E1E1HD$0Ht$(HHHH8HpHhL8HL$HL$HHHt H/JHt HmJHt$1HHH$HHOHD$H@L3E1AI1HD$ DML1AI*k1E1E1AIHD$*:AI11E1HD$hHD$L1E1E1s7L!IH1E1AI/v$8$H$%$ME1AI/=H$$HD${M/AIH1LHD$`HH1AH&Jf.M1AI/H1AHULmbHt}HD$`AH&1HH5$H8 1LHD$TE1E1AI1HH5#H8@yHtH5#H8%hIxA,J1E1A*J78AWAVAUIATUHSHHdH%(HD$8HFyHZwH [H9HL5BMIIFH5"LHHIIHM IHHHD$I9G HL}IMI/IEH5LHHIM{HD$I9F InHM~HEII.LHLHmH HzI/9H+I$HPI$I$HHD$8dH+%(HHL[]A\A]A^A_HzH9wHXH0HqH~1DH;TLHH9uHuH ~H9H L5eM| IIFH5ELHH HIHH IHHHD$H9C HHIHHM HHID$H5qLHH IM H5Eq1LsIH I.ML;=AL;=DL;=LZD$ I/tD$ ID$H5pLHH0 IM/ H5pL9kIFH;Xm I~uA~HHLHHH9AEI.I+HsH@EH9N*L59MIIFH5LHHrIIHMIHIE1H;|$H;=?H5H9t htIOAƃ捁3HHHMtLpHcD$I$Ld@HHEHlIGHH6H=F葼;1HLHD$L\$MH+d I/} L;d L;2  L;@  LL\$L\$I+Z J HqHH9XqL5uMIIFH5nLHHuHIHH~IH HD$H9CLH HH\H+ I,$ IEH53LHH;IMHD$I9F/ MnM" I^IEHI. HLH ImI MGH+ I,$ HEIHPHUfHH9'HuH; fAƉD$LXLHH=)pH*H5+FIME1AO@H b DH=-Mu]MI.LҺ!fA.FH AHIfI,$E1LLM Hf.LXLHEH82L(\LHLLH9i7 L5TM IIFH5LHH HIHH IHID$H5dkLHH| IM IFH; OH;H@hH H@H 1LIM I.?$IH" M{L\$ZL\$HI HmH ;H9H6 L5"M IIFL\$LH5jiHH L\$HH I.H5>kHLL\$L\$:HmfHCHHi H=L\$ L\$ LLHI)ML\$ H+I+I/IEH5LHH0 HHD HD$H9E?LmM2L}IEIHm LLLImH\ H I/H+ID$LLH@pHR H@HE HHT I,$IIHL;AIEHuDLL\$諶L\$L;L;@@LL\$SL\$AA'P11E1E1fDI+Mt I.ME1I/9Ht H+BHHmHT$T$1LIӵfDHoL谵H=jHr}H5s}讁IME1AOgLg%LZHcT$LLt$ Ld$(H)HHl$0Ht IH9 MI.LL\$L\$IFL8ILT$ԴT$H߉T$輴T$L߉T$褴T$]LT$茴T$T˽HE1AOM12E1AO11E1fDH=iI@cIH=hIi@MwMI_IHI/HLHI\I.IL觳fLsMhL{IIH+]HLLL I.IBLW5fI1E11ۺAOfDAO[I_KIAO11E1AO1E1AOM11E1AORMAPEH-@zL-AzHEHH,H=Y1LHIM]LI.VE1AP E1TIA"PHLL~ILL\$ɱL\$lL跱LTL袱H;>L3IHA$P%M~IRLL\$ML\$LHIHL%SHL IH1HL\$L\$LڰLͰ7HL\$L记SH衰;H=eHxH5x|IMAP^H=OeHxH5 xl|IMA2P(L(fLOHL\$ L\$.APE17IH=ds{ILAPE1软H=dG{I M1E1E1AP<ϸHA4P%1HLzIHMA_PH3 L&MwMI_IHI/qH{ID$0I|LЮ1APHcT$Lt$ 1Ld$(H)Hl$0LQHt uI(HAIMSANPM11HK1HHHL訯HmI HAdP11E1E1M1E1AP1E1AP~1E1APLLHL\$辵L\$HI1AP3H=dbHtL\$H5t|yL\$IM1AP1LHLIHA PIAPHIATP{LݬH=aL\$gxL\$IfEHAPUL\$HAPIE11_H=\aHsH5syxIM}AoP5舵HH=awIɺAqPLAP|LsMmL{IIH+rLLLLhI.HGL賫:YMA_PH0HH5T T$H8ȬT$ 躴HAP1E1AP5E1A P"L\$ЯL\$HHHH5 APH8CL\$AFPo脯A PHHIH5 T$H8T$AQL腪ME1E1APH_1HAI%11ۺ1APDAWAVAUATUHHH~IMH^HEH OqH9HL%6qM-I$ID$H5]LHHII$HI$M'HLFIHIFLMH= JLHLAInMI.toI,$tXHmt1Imt9HL]A\A]A^A_fLeHبImuLɨL踨fDL訨I,$uCHFI,$AQYI.D;E1H H=8A4@LHLZIHtI. sE1HH5H|$H|$t蹯IH=9oH:oH5;osIAQMEIH=n,sIAQDLhAQHLIHH5H8~fLAWAVAUATUHHHnIMH[HEH 'nH9HL%nM-I$ID$H5 ZLHHII$HI$M'HL6IHIFLMH=:LHLAI^MI.toI,$tXHmt1Imt9HL]A\A]A^A_fLإeHȥImuL蹥L訥fDL蘥I,$u3HFI,$ARYI.DBE1H H=C1~4@LHLJIHtI. sE1HH5H|$|H|$t詬IH=)lHlH5lpIARMEIH=kpIARDLXARHL9HH5H8nfLUHGHukHSH}HtHEH/tfGHD$Ll$xHD$XHD$`H$HD$hL HD$H!HjT L(hE1ɹAHƺLAIHH8FI^H5jLHD$XH\$0[HD$HHpRf/, L9}踟HD$8H f.HL(M9t MHPHuL`H@HD$(MtIEMtI$HD$(HtHHuUH iH9HL=iMIL|$XIGH5QLHHILL$hMiI/hHD$XHI9AH|$hHHD$`IHD$XH|$hMTH/8H|$HD$hHHӣHD$HD$XHLxHX HD$`HD$hHD$XMt ImMt I,$lHL$(HtHHD$8HH@fDHTH jhH9H@H-QhHHEHl$`HEH5OHHHHH\$hHHmHD$`IHHD$HID$ HD$XIHHH5UQH HCHHH==LLHHbHH+ HD$hI,${ HD$`I/X HEuHD$XH} HEHD$8HPfH=uXH*D$HD$(KXHD$ L%PLLM}L`HHH@HH@ HLLHHEHD$ L-5PLLM|$LHHH@HH LLHD$XHHH@H;j LgL|$XLd$`M LGI$ILD$XH/T LL.I,$IsHD$`MFH|$XH/ HD$XI/ ǣHD$@HD$(HH|$HD$(HHD$ L|$8E1Lt$HLt$0Hl$ LH`Lh HD$H\$8LMMH HHL$Ht$HMILLI賫H\$L9|$(uHl$ H\$8Lt$HH|$@莚HCL%NLMH=,wa 1LHAHD$蚠LL$M0 H+hI)HEI. HmIRE1H#HGWHH HD$ fGHHD$H|$cH2EGWHH HHD$fHD$1E11L|$XEMt I/E1H|$`Ht H/Ht H+H|$hHt H/Mt I)(H bH=ZrMuhHt HmMt I/H|$tHL$HHD$HHH8Hm-H蓙 fDI.nE1fH|$覜HD$XIHx{HD$HD$`HLxHD$XHD$`DL`H@HD$(efDHDLLωt$T$Иt$T$Lt$ E1T$LL$記LL$T$t$ HfDt$ T$LL$~t$ T$LL$2@H߉t$ T$LL$St$ T$LL$t$ T$LL$.t$ T$LL$HD$IDII iM I&HFHHD$pIH5kaLHVߜHD$xHIM H|$pxf.IHF(HH$HF HD$xHFHD$p蠕IM~1HL$pMLL H5mGy&Ef.HY1LIfDH1HIEH9t5XI*GAFIYH\$XHcIyHHH|$hI)t4f.HHHD$`IH+:H-LH\$XHH|$hH5j?LHVHH$IoHuHD$XM HHH@Io HIW(PHH&E HU0HMH@HDHT$E1HcH\$(1Ld$0LIA2Hz0LRH@LED9tY1HyLHHw ITz rLBMtMIJ  uLRHD9uILHD$IH<0LyLf5KAsE1E1H|$HHt H/H|$PHt H/H|$XHt H/Mt ImH|$`HtH/tdH FDH=iMMt E1I/t.MHD$hdH+%(,HxL[]A\A]A^A_DLMtt땐{tjfDLhtg[t"fDKt&fDHH5'LHILl$PMI9L;-u L;-ImHD$PsHHD$I9 IT$ID$H5<LHD$HHMHHl$PHJHEH;H;H@hHH@H1HILt$HMHmjHD$PIFInHEH HHcH>f.AnHcH I.ID$H5%LHHILd$HMID$H5U;LHHILt$PMI,$HD$HIFMfID$H HXHcH>fDAnAFHH H"fAnAFHH @AnEfAMcI I.H&H d:HD$PH9HIL-B:MIELl$PIEH5\$LHHILt$HMImLktHD$PIHj@uHD$XIHLxvwHD$PIH5H%H 9H9H L=v9MIIGLT$LH5#HHYLT$IL\$`MI/t H5U#LLL\$ LT$x L\$ LT$I+IFLMH=LT$oLT$LT$LLLAI1uMLT$fL|$`I.HD$HImHD$XI*HD$PHD$`L;|$iH$H5"MwLHt$IRLT$(HHT$ OuHt$HI H@HT$ LT$(HHpLLIH&H$H5"LHt$IRLT$(HHT$ tHt$HH3H@HT$ LT$(HH LHD$PHH'H@H;7nL_L\$XMLWIILT$PH/<LLL\$L\$HD$`II+%HD$XM]H|$PH/LHD$PI*)HD$`lHL$HDHLMt I9l HPHuHHH@HL$HD$ MtIHD$HtHHD$ HtHHKIHH$L\$0HL|$8HH\$H Ll$(IH$H<$LmHL$LLHL<LqHLLsqHLLeqH)IuLl$(L\$0L|$8HD$H8HD$`HD$`Mt I+H\$HtHH$HHfH\$ HtHH$HHUIEH-HHaH=#nk1HLHqHIm>HmHD$HIX@EfcEfAFII I;fEfAFII (@HEL0ILt$HHmHkLkLk?LkLuIfDLhk0LXkHD$HH\$(Ld$0Hl$HI/uL-kSoHD$`IHHhH3HIG zqHD$HHHHZH5S3HrCHD$H@LMH=[iH|$HLAIoMLl$XHL$HHD$HHHD$PI/HD$`HmHD$HImHD$XH$L5!HLLmLoIHH@HHLLHIH"H$L5HLLmLoHHH@HHK LHHD$HHHu!H@HD$`H; LoLl$`M LGIEILD$HH/9 LL跹HD$XHImHD$`HTH|$HH/ HD$HHm HD$XgHD$HHL(Mt L;-HPHuHHH@HL$HD$MtIEHD$HtHHD$HtHHHzH$L|$(L5ڸLl$ MH H$H<$HgHIEL9NH;H@hH H@H HLIL|$XMIEL9QH;jH@hHH@HHLILd$HM IEL9tH@hH H@(H LHLЅI/HD$XIEL9H@hH H@(H HLLЉŅI,$HD$HHLl$ L|$(MtLSH\$HtHH$HHH\$HtHH$HHH51LF3I/IqMImDL YIMzDKAE1E1LjH;fLfLL\$ LT$eL\$ LT$kL=MIT$I9HXHHqH~1L;|HH9uIOHRaKE1HMH5A{E1H81nfDnHkKA|H|$HE1E1HLiIF@iH%pKA|E1E1@iHKA}E1E1lffA.GzfHH-E1HH9IA@ kIo HIW(^nHH9 LjŅ/TKAwE1E1HH|$PLfLc_IH`H5E1aKA{H8effDLc 1afIH^HHdImI,LdcL@EHqHHt$ HT$>aHT$Ht$ j@HmuH cHD$HNAE1E1f.{KA}lI$IULwYHuHKH5H8UImHLLHLAE1T9|TBH;L UHD$PIH uLAH;}8 LHHl$HH NALSIo HIW(]HHH.LQNH8SHD$HH|$P_NtWHxXHX`LpXLxhL``HhhHt H/ Ht H+ Mt I/ HD$HNHD$XHD$`E1NALAE1S,NAH|$HLZHD$XIHQE17NApHDHI9HuL;=̢E1LpRLcRLVRhLIRbID$JH8IL0H/;"R1LAX[HLAE1E1LQ]NcQID$HH8IL8H/Q LQ0HHH@HL$HD$yQH|$PLD$`IMNHD$XI1HLdYIQH;ݠH|$QH|$HHD$XI E1LAN}IELPLPLPgUHuH;H5tH8QHD$X7NE1E1A5LuPLl$ L|$(NHD$XYLKPHH|$PTH0Pw&P謥IL P3LOOTHbH$HgH5H8QL$Ll$ Lt$(`MAHl$0H DH=$HD$Xv(H|$HL$`HT$PHt$X="HD$XLd$`HT$PHD$ HL1H$[YHD$HHHt1HLImIxH+WHD$HM3I9L;֞L;LLT$(TLT$(I*y8MHD$HL$LE1HT$AH蠪3MͅHD$ HtHH$HtHMtLHD$HL$LHT$H>H'HD$I*uLMGLPH LHLHD$8KLL$8AI)LMLl$ Lt$(bMAHl$0H|$XHH/4$zM4$LjMH-PIHKLHL3KI.AL.MH=H{H5|/IL|$`M[E1LAH=ILl$ Lt$(nMAHl$0!Ll$ Lt$(lMAHl$0E1LAAUILtLLgLqLZLQE1LALA}UIH MAH8KLLl$ Lt$(xMAHl$0ME1LAH]H5ӱKAH8MfHLT$(KLT$(LHD$(KLT$(qIVHT$HH\IFHLHHD$PLt$HM5H|$PLLיHD$XII.(LKLASTIcLAH=H]H5^IL|$`M E1LAeSI#H=4IH=xHH5IL|$XMt,MHNJLAJE1LAtSIH=IE1LALI&NH$Ht$ LHǽMMHD$XHD$PHD$`MLAE1'JHMH8IHD$PE1LAHHI9HuL;-bL.IL 7IMXH|$PH I_HHpH;H5LGHE1LAxMMIE1NALH-H9AL;=lE L;=| ImT$tMHD$H|$L NAƅKKAwE1E1KALH멽 MAQNH9AL;=ʗAEA L;=ؗA H/t;HD$PELMAƅLAE1E19|GE1MAIEHH-kH9AL;-2E L;-B I/T$uLGHD$X|$MLLAƅ:xLAE1E1KFM"E1ONAoIx\HcHA9T|F~J1fD}6p9}%)HcHAL9~؉9|9AH1H~?AUIATIUHS1HfH}UADHI9uH[]A\A]fDHHH?$L$PYD$X$H@LVM1HI9tH9|ufE1DJTHBtv@tmH9tHXHt,LAM~S1HI9t?H;TufDHDHH9tHuH;Prf.IM9k1AVIAUATIUHSHC1LLHICHtpH I1LH~HTHHHH9u1L8FIAE PAU I,$tAE [L]A\A]A^LXDAE E1f.HHH8Pf(\zFfWBHff.fHD$YD$HfSHH$L$ff/wFH;Sf/r \\f(EL$Y$H[\@XEYD$X$H[fDAVfI~SHHL$H;Sf(\f/vgEfW/ZEL$H[YfInA^\ff.@SHH$L$fH;Sff/v x\^DYD$X$H[f.SHH Gt)GGHGH [ff.zfudHH8PHXH8f(\T$PT$Xf(Yf(\ Ψf(YXf/sfDf(L$T$\$D\$fT$YL$^f.wQYCYSH [L$T$JL$T$H$L$YD$X$HUHHHD$L$H]^f(ff.ff(f(SHXf(H ^L$\$l$9\$L$Y f(YYYf(YXff.w_Q\YT$H\$H8XT$P\$T$f(f(X^f/s Y^f(H f([f(\$T$kI\$T$f({fDf(f(H8H\\T$H?L$D$(f(d$^l$ 4$P4$d$L$T$f/r7l$ f\$(YYf.wXf(QXH8f(@f(\Yf(\Yf(ff.w&Q\H8f($gH$f($RH$ff.UHH0f.D$z^|$ff.zfty=pf/|$@HEH8PHD$( @t$T$\f/r: "D$f(^_>\$f(f/rH0f(]DD$L$ \^D$M@|$L$ D$f(Y\f( ^=T$\$f(\f/H0f(]f|$\=Y|$(ff.Q56^t$@Hff(D$YXf/sf(HEL$YH8YD$PL$pf(YYY\f/wbL$ #?D$D$?%\d$f(L$ f(XYD$(YYXf/D$$L$(YL$H0]f(fDH0]Eff.HL$MYD$HfHYP+HXfUHH $D$ $Hf(Yf($\$$H]Y^f(fUHHD$ L$HY $f(L$L$ff(f.w!QY $f.w1Q^H]f(D$f(DT$ff(f( $bD $f(f(ff.zuHT$y T$H^:ff.HD$=^D$D\JHDHD$ fWuD f(^L$H\f(M:ff.fHHZDf.AVUHSH D$f/L$f/fHEH8PHEfH~H8P o^L$fI~fHn9 R^L$$fIn9X$=0f/rf/zv` $H []^A^f(f.D$HBH$D$/ $H []XA^^f(DfHn6;f(fIn^L$ $; $f(^T$T$f(_\\$f(L$B\$T$$\f(BX$:L$H []\A^f(B@SHH05 D$ \f(y:D$(H;Sf/D$ D$H;SYD$(JBžT$\f(Yf/~f(T$\$:\$D$f(9L$^jX?L,M]T$ff.E„AH0L[f/ArA@SH@f.D$8L$ IHf/tl$ f/f|$D$^XD$0-@f(^f(9XD$L$\f/D$syH;SY,9\$0H;f(YXXL$f(L$ ^\d$(YL$SL$\Y\f/T$^H;SD$D$(/9 Gf/L$vfWT$8 yXf(fTT$XK&8t$T$\2f/v.fWH@[fDH?SX\VYH@[t$  f|$YY5t$Xf.wmQXL$|$f(Xf.w^Q\|$ f(f(X^f(YXXL$^L$0<HMf(">f(f(L$ >L$f(닐HGHHt f.=ff.AUIATLHH=oHGHHtIMtHLA\A]Ðk=IMuHLH5H819uIHHH21Au*1!H~H9ut1AH[]A\A]@I9uuHDL/HHH;f~H;-4~uH9u8HmuHD$ ".D$ @1AH[]A\A]@H3fDHM0HH@HE@+H:L+9fDIt$HfDM>@M>ff.H;=}}H;=K}u H;=]}u 3ff.AUIATIUHSHH(dH%(HD$HGXHGXHHt$HD$HG`HG`HD$HGhHGhH|$H$-H{XH4$H|$Ht+H4$HtHH|$HT$HtHHtHHIUI<$HuLL`HhHHxHpMtI(tAMtI,$tEHtHmtI1HT$dH+%(H([]A\A]L+Mu뼐L+Hu븐H+fDH|$IEI$HEHtH/t0H|$HtH/t8H<$HtH/tXs+ɐk+f[+4,@HHHH H5(HEHzH813ff.ATUSH)HxXHu1H[]A\HHN{H0H9ufLc`HkhHCXHC`HChH/t0MtI,$u L*@HtHmuH*f{*fHGta@tXHFHt_@tVHXHtXHJ1H~H;t?HH9uH|$8+H|$ݩtHHH9HuH;5y@ATUSHHf(HxXHHu1H[]A\HyH0H9Lc`HkhHCXHC`HChH/twMtI,$u L+)HtHmuH)fH/t*HxHH5H811H[]A\ÐHt$(Ht$(HGt_@tVHFHt]@tTHXHtVHJ1H~fDH;tHH9uH|$)H|$ݩtmHHH9HuH;5Lx@AVIAUAATUSHH dH%(HD$1Ld$Hl$HD$HD$1LHHh.tTHD$H@uHkwLH5H81'01HT$dH+%(uKH []A\A]A^DEuHL$HtHwLH5pH81/1'DAWAVAUIATAUHSH(%H=IHH@XMw`IGXIG`H$IGhIGhHD$-HHH8HKH9G)L 2MHbvI9tL; ~vL(E1L EAMJDDLDeA9,HHLD;pIH(HEH1HL$IHXdH'HmuH%I,$LH([]A\A]A^A_5)DHyuH5E1H=*H$IXMW`MOhMw`IGXHD$IGhHt H/Mt I*,Mt I)5L EM|DLDHD9_ADHH1H="IHH+HHL.I,$HHL EEMKD$DLDAA9HcHLD;vD9AJDHcHLoHcƒHHAA9|ADvH.DHEHm$HH([]A\A]A^A_#EHL-HfAD9EEj@LIcHh%IHkIcDHHD-HE9 ;LL$L$#LL$L$@LL $"L $L"H=H5HGHHLHHH|$6%H|$HHHHrrH9r Lh"H5iHL$HVk'HL$HIHH@Hf+HH@HH5DpH(HE@IcHLHHtHqIHD$LL$!LL$HD$*H>H.H/!L qHdq*HDATHGIHHt.HqHЯH81+u.LA\HpHIHH5AH81)I,$tE1LA\L fDUHHGHGHHHMHcH>1H]GH]WGHH HcʉH9tHRpH5H8!@WGHH HHcʉH9uH]GH]$HcH9fHu$HtLf.H@`HtkHHt_HHtUHoH9Eu,@HHmHD$ D$ HH5 HHux$HjH'oH5/H8 OHGtgHGHPHwQH ίHcH>G@GWHH fGWHH HGH#UHH@`HtvHHtjHHt`HnH9EuHHHmt)H]HH5HHuHHHD$;HD$"HuHnH5 H8dfATHHGHLgID$HwuHHcH>DgAMcfDH/t HLA\HLA\ÐDgf.DgGII DgGII IH|$3"H|$IfH@`HtGHHt;HHt1H@H;nmu:&H|$H|$I@!Ht&I0H5~HHtH@HlH5H8 HHH?Pf *YHHGH?H~OAUIATIUHS1HfI}AUf *YAHH9uH[]A\A]HHH?PHHHHH?PHHHH?PHHHGH?f.tE„af.f(D„GUSH(-f/f1XH^ HP^YHYXH9uf(d$L$T$@T$L$f(5d$^f(\f/YX bX\vFH~AfD\`L$Hf(T$L$H9T$\}H(f([]f(f\H,H*Dff(AWAVAUIATUHSHHxHT$PH E1HHxL[]A\A]A^A_@HD$PII)IH9!fH*fI~fNd5fIn\$VDH;L$SfL$I*f(If(^Xd!L$H,fH*\M9tf/L$wfIn\L,L)L;l$PLO2I9HfHINffIMMdH*MLII*ffI)I9H*|$LOf(^I*L)f(YXD$fH*IT$Y%^\YfH*Y^Xf."f(Qf(HHIWIYfIHD$0XT$`M)H*L$XHT$(D$ fH*\$hHL$8YfI*^L,fIL$H*HL$8D$fL)H*\$HT$(XL)fMH*\$XD$Lt$HD$fI*d$HD$0L$XT$`XL9d$@Y D$ ~%X}Xf(fTf.{]T$0fH;SH;D$SL$|$\ YD$ ^XD$f/wf/D$0sML,fID$M)H*HD$8D$(fL)H*d$(IFXfH*d$(XD$(HD$HLD$(fH*yXD$(T$@L$\c\Y\f/s=f(T$\Yf/|f(T$Xf/L;l$PMOM)L9MOfH*fI~DH,f5G|fUH*f(fT\fVf(M\$hf(HD$T$(\$ T$(\$ HD$f(AWAVATUSHH@f/D$|$f.=zAuH@L[]A\A^A_ÐD$fWE1 r{D$f.IL$H;SL$Yf/L$wH@L[]A\A^A_fDf(ff.)Qf(D$L$}L$Y X D$0Yf(L$ \ \-3f(\%^ f(d$8XfI~XfI~^\fH~fDH;SH;f(\T$ST$ D$f(fT2z\fIn^L$XD$ YXD$XW"L$f/ LL,\$rfHnf/'M_5'f/v f/Gf(L$(D$fInL$(t$8D$D$ Y^X|$ID$X|$f(f(\fI*YL$0\L$D$fH*L$L$\f/L$H@L[]A\A^A_D$rf(fUfHf(Hf.zxf/wbf(H} $YT$fT$HHH*Xt $fH~f.HJH]fHn\HL$f(:H$L$ff(f.w+QXf(YX$H]fDH]D$f(_T$f(UHH $f(D$$$Hf(Yf($\$$H]Y^f(DTwSH\^f(/H;[VfDGAWfAVH*AUIATIUSHH$t H9r vMe$AEf(\Aed$Pf/l$H/ L$P|$HT$YAMA} X\$AM(f(L$\$Y\$HH,fL$T$f.Im0f(\$h Qf(Y\$HY%v-\f(fTf.A=IXf(|$(f(D$AE8ff(H*XX*Xf(Au@$\f(\$xA]Pd$8^f(AeH\XfD(\$@A]X\$PYf(f(\^f(YXYf(Yt$H\\$`^A]`Yf(XYfA(AXXL$pAMhYf(D$AEpfA(^D^Xf(D$XAExfA(XD$ AMI)IEH$@H;SL$ H;YL$SL$f/L$f(ff/L$"|$l$@fH*\Yf( t^XL$0XD$8\X\$(fTs^\f/T$[T$L$0L,MI)LHIHH~#D$hYD$(fH*\f/ID$fd$P^d$HH*YI9=f(sf/$M)f/d$(MGH[]LA\A]A^A_fDf.XADEf/L$XwkL$0D$Z^D$`XD$8L,MSEJL$0\L$T$Y rYT$`fL$0D$ \$x^D$p\f(L,M9ET$L$0\L$XY !rYT$pifDHEf(I9ff(H*H^\YI9}IFH9ff(H*H^\^H9~Vff(^ LfX t$hIf(YXX ^HH*^^f(XL$(YL$0\$ \$L$0f(f(\f/XT$0f/IFffH*fEHEH*$fL*ID$L)H*fD($$DYfD(DYfA(fA(D$^AYfD(l$DYD$D$$D$ D$$D$fA(^D$ |$Y|$P$\$Hf($Y^S 5{$-rfD(D$D$f(D$^f(D$D$DY$D$$\%f(^\fD(f(^D\D^ f(A\fEM*DXL$(DY$f(T$0A^DEXfEM*EYfD(EXA^AXfD(E^E\fD(E^E\fD(E^E\fD(E^fD(E\E^E^AXfD(E^A^E\fD(E^\A^E\fD(E^\A^E\fD(E^\A^E\D^\^L$A^E^DXDXfA/u$YT$\XL,qf.H,ffUH*f(fT\fVf(f(l$P\t$Hf.Blfb8jz r`d$b@YHj0t$`rp$bHl$P-hd$8bPYt$d$xbX|$Hd$@bht$ d$pbx\$hd$Xl$(D$hT$L$T$L$Rff.@AUIATIUSHH8$t H9rL$='lfMeI*AE\A]A} f(T$|$mT$YT$X$T$L$D$YAEfYA]XX kf.Q%f(YXf/-H,Im0H;Sd$E1f/f(wWH;SL$E1f/vAI@H9|Lf\L)IH*Y$YfH*YT$^f/wH8L[]A\A]@f.Br Hj0t$rt$Lf(\$(T$ L$- %\$(T$ L$YXf/f(d$ \$ \$d$ f(@YXH,fDfUHH*f/r!Yf/r_]fi\Yf/f(riIH]L)fDIH]L)]zf.Hf.EunfUHH*f/rY'f/rI]0i\Yf/f(r&IH]L)fD1D]fIH]L)HhHH?D$\$PL$$f/vf(fDYHXf/wHÐAVSHH(\jhf(D$D$DH;SH;D$S%.h\d$fI~ h^L$f(\=f(hfTf.v;H,f=gH*f(fT\ gfUf(fVf/]5gf/Kf(L$T$^XT$fInl$f(\YgY^Yf(\Ag^f/H(H,[A^fD1HATIUHSHHH HHH HHH HHH HHH HH H HH uI<$AT$!H9r[]A\I<$AT$H!H9rDff.@AWAVAUATIUSHHtpHHH?IIH9wmHEAEDjAME9v(D1AAA9sH}UAME9wI IHL[]A\A]A^A_@HEHEuQLbIHHI9v+L1HIIH9sfDH}UIHI9wHL$DHEH}L!I9rL$oDHEH}D!D9wL$ODЉI>@I0fDAWAVAUATAUSHtiHGIH?AEudJL$ Dl$ AME9v)1AAA9sDI>AVAME9wI AHD[]A\A]A^A_@IFI>D!9rD$fAfAWAVAUATUSHt$Lt$PfAIALf$AEDjH?AT$AAAfA9AAǙAf9r1@t$ I<$AT$t$ AAfA9vptA.I<$AT$AAD!f9stA.AD!f9rfD$H[]A\A]A^A_D$H[]A\A]A^A_A.AtA.AD$fADH?AT$Aff.AWAVAUATUSHt$ Ld$PAIALˀ&AEDrH?AUA$DA$$A8AAAǙ8r-I}AUA$DA$$@8votA,$DI}AUA$A$D!@8stA,$A$D!@8rD$ H[]A\A]A^A_fDD$ fH[]A\A]A^A_fDA,$A tA,$AD$ A$@H?AUA$ff.USHHl$ tAHL˅tmAEH[]fDH?QEAWAVAUIATIUSH(LL$Hu.LIH~fL#HH9uH([]A\A]A^A_DHHH9FEH~ĉDrE1Dt$D$@H;St$I9s$D$19sfDH;SI9wHD$H LJ IM9uVf.HEHLrE1Hl$H!DH;SIHHI9v-HD$1IHH9sf.H;SIHH9wHt$HLJIM9ufH1DH;SHt$LHHI9uH1DH;SHt$LHHI9uZfDIII LHI LHI LHI LHI LH I HE1H;SL!H9rHt$LJIM9uIII LHI LHI LHI LHI HE1H;SD!9wHt$LJIM9uufAWAVAUIATAUSHL $u0LIH~@D#HH9uH[]A\A]A^A_DHՃEu}DrE1Dt$l$ H~fDH;St$I9v$D$ 19sfDH;SI9wH$H DB IM9uH[]A\A]A^A_ÐIII LHI LHI LHI LHA H E1H;SD!9rH$DBIM9uH[]A\A]A^A_H1DH;SH$DHI9uH[]A\A]A^A_AWAVAUAATUSLHfu1IIH~fD+HH9uH[]A\A]A^A_@IfCEH~DpI4I1H4$A1D$ uz@I?AWfD9s@D$ Af9s1t1fA9vI?AWȉfA9wۿHDfKH;$;t1뒐AHHH HHH HHH HH HM4I11t@1!fA9sI?AW!fA9rDHfSI9uH[]A\A]A^A_@HI,I11'DI?AWALHfKH9ct1ff.AWAVIAUATUSH@t$ uRH~8H@HL[]A\A]A^A_@f@t$ H@sH9uUH[]A\A]A^A_fDIEH~I, DzLˉD$ 11Af.8I}AUDA8nD+D$ EϙAA@8rL$HcT$\$L$%D$Yf(YUT$L$f($YD$Xf/wAI>AVfII LH!H*AYLtfW I94?H(f([]A\A]A^A_@IFI>fWb 2I>Y $AVfWr= $fW`f(XYf/vX AzfW 1mff.H~?AUIATIUHS1HfL8ADHH9uH[]A\A]DUHH0f.lKD$z^|$ff.zftv=@Kf/|$@H}UHD$ Kt$T$\f/r= JD$f(^2\$f(f/rH0f(]D$JL$ \^D$|$L$ D$f(Y\f( sJ^T$\$f(\f/H0f(]f|$\=¼¼Y|$(ff.Q5J^t$@Hff(D$YXIf/sf(L$H}YYD$UL$Cf(YYYI\f/wbL$ D$D$%]I\d$f(L$ f(XYD$(YYXf/D$'L$(YL$H0]f(fH0]ff.HL$MYD$HfHUH\^f(H]fDUHH -|HD$f/L$rf/sFD$HH$D$ $H ]X^f(@ff/wnH}UH}$U$ G^L$D$f(9T$ G^L$$f(X$=Gf/rf $H ]^f(ff.fHYHXfUHH $D$ $Hf(Yf($\$$H]Y^f(fH$L$YD$X$HHHJf.UHH?HD$1L$H]^f(ff.fUHHD$L$HY ȸ$f(L$L$ff(f.w!QY $f.w1Q^H]f(D$f(T$ff(f( $r $f(f.UfHf(Hf.zEf/vX\HL$f(H$L$ff(f.wzQXf(YX$H]fDY H$f(Gf$HHHH*]XuDH]ffDD$f(hT$f(lfUHH $f(D$$$Hf(Yf($\$$H]Y^f(Df(f(SHXf(H ^L$\$l$\$L$Y f(YYYf(YXff.w\Q\YT$\$H;XT$S\$T$f(f(X^f/s Y^f(H f([f(\$T$\$T$f(ff.UH@f.D$8L$ gHf/,ܵf/D$  l$ f/  =BY|$Yfl$Xf.tQXL$t$f(Xf.dQ\l$ f(f(X^f(YXXL$^L$0-@f(^f(XD$L$\f/D$s{H}UY\$0H}f(YXXL$f(L$ ^\d$(YL$UL$T\Y\f/T$\H}UD$D$( f/L$vfWT$8 FXf(fTAT$X|$T$\f/vfWH@]50Af(t$t$ ^Xft$D$0H?UX\@YH@]f@^D$ f.weQD$HFYD$ XD$8f/vXTf/<E\>H@]"D$f(f(~f(L$L$f(fDSHH0D$ fWV!D$(H;Sf/D$ D$H;SYD$(R?T$\f(Yf/~f(T$\$\$D$f(L$^r?XL,M]T$ff.E„AH0L[f/ArAff.HD$~ET$$fWf(~% $f(fWf(^kHH,Ðf/4r DfT fT<>fVGGT@Ա(T.v,,f%U*(T\V(@(T.v,,f(%U*TXV(@4p(T.vXU\V(ff.@ 0(T.v,fU*V(ff.f.wQifW'W7wgAVAUIATL%܉UH-ԑSH̍HI}AUfɉ *YL9IEI}AL$HcAT$ \D$L$A(W-?A (#T$ L$(fA*YcYD$X/HH([]A\A]A^f *Y+WԮ #H[]\A\A]A^(H~?AUIATIUHS1HfLADHH9uH[]A\A]DAWAVIAUATL%ZlUH-RhSHJdHIFI> HcL$ \D$%fAnfZAYA f(YL$ f(fA*YYD$XZf/wDI>AVfAA A*AYfA~t WofA~D9l/HfAn[]A\A]A^A_IFI>f *Y|W%0 xI>YL$AVf *YDWL$W۬(XY/vX 'AfA~RW fA~AH~?AUIATIUHS1HfLADHH9uH[]A\A]DUHH .D$zd$f.z f=f/|$%Qd$DH}UfH *YT$T$  t$T$ \/r5 D$ (^\$ (/rH (]@D$«L$\^D$|$L$D$ (Y\( ^T$ \$(\/H (]t$\5RfJYt$.5Q=%^d$|$H8f(D$YXת/s(L$H}YYD$ UL$f (*YYD$YY~\/w_L$D$D$ =Q\|$ (L$(XOYD$YYX/D$L$YL$ H ](fH ]aff.HL$ YD$ HfgT TVf.g.(z$f.z(tf/wݨ@YYԨHSYHfDYff.@.(zt"(fH\/w/sIH@X`(@WɧT$ NT$ HXf.L$ %L$ HXÐ.(zt"(fH\/w/sIH@X(@W9T$ T$ HXf.L$ L$ HXÐf(f(f.z&ff.zt"ff/w`2ff.@YTYLH3Y;HfDY,ff.@f.f(zt(f(fH\f/w"f/sDHXإf(fWT$%T$HXÐL$L$HXÐf.f(zt(f(fH\f/w"f/sDHX1f(fWPT$eT$HXÐL$EL$HXÐl$z,zl$tw @DXff.l$-ff.@l$-Ҥff.@Ht$t$ Hfl$|${ff.Hl$ l$ztwsPHÐHD|$H<$ H<$~l$0H H|$ H<$H<$Jl$@H Hff.Hl$ l$ztwsPHÐHD|$H<$H<$l$0H H|$ H<$H<$zl$@H Hff.AuA1ADuDDADÐATASHAtD1AAHD[A\ff.IHHu I1ILHuL@ILfATISHHIHtL1IHIHL[A\IHHu I1ILHuL@ILfATISHHIHtL1IHIHL[A\AAAHAHuDIHIIHHIHIHIIHHIH#AAAHAHEDIHIIHHIHIHIIHHIH@w@ljD1D@w@ljD1D@w@ljD1D@w@ffwljfD1DfwljfD1DfwljfD1Dfw@fÐ1 Bff.@1 Bff.@1 Bff.@Gff.H1HH@HBff.H1HH@HBff.H1HH@HBff.HHH?HH?HGfDH1HH@HBff.H1HH@HBff.H1HH@HBff.HHH?HH?HGfD(T B.  v (Df~f~(%=#u-D$L$. ȝzNuLL$\(fDBt5fnȁ(YT$T$@ pfDXff(fT )f. P vf(ffH~fH~H ȉ%=~Z-f( кu;%==f(YL$L$H H fHn\f(@ uDHD$L$f. zuL$\DHfHn\fDgq%=df(XH8dH%(HD$(1l$@-|v$HD$(dH+%( H8DHD$Hl$@|$l$@z?u=HD$D$l$<$,$zu,$l$@wDHL$@qt$H L$uP% ‰T$l$tIu<$,$HT$@HH ul$@z?D$l$Yfl$釿SHHJHtD$H[ fDHdH%(HD$1H|$D$HT$dH+%(uH߿ff.@ATeAąuDA\6DA\HdH%(HD$1H|$D$HT$dH+%(uH_ff.@ffffH(l$0l$@-vw6@zrz t.@|$@|$0H(5Dz, z|$<$l,$l$|$<$Tl$,$|$<$<,$l$vH~P$f(|fTf.v fTf.w f.zjf. jztH长Df.z*f.zL$$L$$H`L$$$L$L$$L$$vH(T._v T.Sw.zp. ztH,@.z3.zL$ D$ L$ D$HL$ D$D$L$ 뮐L$ D$L$ D$nff.S(HH.h(L$\$L$f\$(.z(.\^z//8t X\%.z/u-^(T$(xT$H[(T$d$d$T$\/%vXH[Ð(\$ d$L$\$ d$fL$(T.ݺ(^E„Q.HD$T$ T$D$&L$D$L$\$vff.HdH%(HD$1. Azt-H|$#D$HD$dH+%(u&HHD$dH+%(u H袺fH(fdH%(HD$1.zt-H|$(HD$dH+%(uU(H((.ú^ET$ u.zT$ T$ fDSf(HH f.vf(L$\$L$f\$f(f.zf(f.\^zf/f/8t X\%{f.z-u+f(T$^f(T$H [@f(T$d$Sd$T$\f/%vXH [f(\$d$L$\$d$fL$f(KfDf.ݺf(^E„Kf.AD$T$6T$D$DL$D$?L$\$hff.HdH%(HD$1f. xzt,H$HD$dH+%(u(Hf.HD$dH+%(u HҷfH(fdH%(HD$1f.zt,H|$f(HD$dH+%(uSf(H(Df(f.ú^ET$uf.zT$@#T$6fDSHH l$@l$0|H<$|$Pt$Ht$HH l$@z l$0z @8tzDuH<$l$0H |$<$l$ H  f.;H [|$H<$|$l$XZl$v;H [<$H |$|$`<$EH ,$l$@Al$0l$0E„OU|$<$Yl$,$?Dql$@tH(dH%(HD$1l$0l$@zt6HH |$<$l$ H HD$dH+%(u*H(@HD$dH+%(u|$@|$0H(fH8dH%(HD$(1l$PztDfDH|$H<$t$Xt$XwH HD$(dH+%(uGH8fDl$@<$E„uz,$fD1,$EDAUATUHHH@ubtk@tb1[IHt<1HH6ImItlMt IL$@tiLH.I,$t7]A\A]ÐH]A\A]H]H5BA\A]H8D]LA\A]飲L蘲fDHiHH5BH81%AWAVAUIATIUSHHLw0LHo8H==AVuA~$CtVHKHHS@HtHEHHEE1LQLLR1PUjaH0I%HL[]A\A]A^A_Ht+HEMu IcVH9t/HKHHS@E1IcVL9t!HKHHS@yHuLLlILLLxlIvAUIATUHSHH(HdH%(HD$1H;=NH5H9t wtHUBፁ3!IHIELhHHX HEHEHHH=d误1LHIյMtXI,$t9Hmt"HD$dH+%( H(L[]A\A]HXfDLHHmuIHE1I,$uH4$HHHT$IuE1f1LHI;L,$1H\$LB tHu7AIfDH}HH5TH8̰1ɺAI@AVAUATUHHHH;=cHt$L%L9L耷u|Ld$=IHI$L`HELMTH=حp1LHAIMdImuRL襮HHUBLt$LjE1 t9H=*uujLLAI蠳MtCHL]A\A]A^@LefHt$HHI]LA\A]A^@òIHHE1]LA\A]A^DH}L9tL?HUBፁLB1 uH}Ht$|AI9@1LH裵IHH5=H8赮E1IHuHH5H8}t1ɺAIfDATUHSHH;=H5}H9thCu_HEL%aHHH=m 踫u)1LHIMuWDCIHtkE1[]LA\HUBtHZE1 t2H= ^u1LI英MtL[]A\fDLef[H1]1A\HH5 H8Z[LH1]A\ff.@ATUHSHHGH;fttH;mt;HXhHHCHt HHH[]A\Ht HyHGtH;EsEHDHH[]A\Ht HyHGtH9EvHUHHH[]A\HIHt`HH腬I,$uLHD$HD$f.HH@Ht$Ht$HxHHC#1HHt$H8襬Ht$tHt$"HCHt$@ATUSHHWHBhHHxHHHH9FHFHPHw3H:vHtAHH[]A\HvCHH fHHurHHtHHH2tL;HCH59HPHH81+%fHIHRH5~9H81H1[]A\fHxIHeH蔬I,$HILHD$ Ht$2Ht"HZHf1fvCHH ff.HGH@pHtH@Ht@ff.AWAVAUATIUHSJHHHL$L|$(Lt$ LL$Ll$0dH%(HD$81HD$ HD$(HD$0LLLLϮH Ht$ HHHfDHPHHtH92uHT$(H|$H)HHFfH%f.HFH9GHJHHt5H9H9uHL$(H|$HH)H:H ?Ht$ f.H9u&fDHFH9GHH9HEH8H9uHT$HH57HH81詯HT$8dH+%(HH[]A\A]A^A_ÐHT$HT$x:Ht$ DSHuHt$ HT$ʬxdHt$ bHH9IHHT$H52PHFH"HT$H5V2H81ܮ.ͪHHt$ DH9+HATUHSHHFHHW@@HXHt7HJH~i1f.HH9tRH;luH[]A\fHH9tHu1H;-vH|$gH|$uIL9uXH1[]A\f.HH[]A\0H^H~1 HH9tH;|ukfE1JtH9OHGm@`HFHtR@tIHXHtRHJ1H 1. The Zipf distribution (also known as the zeta distribution) is a continuous probability distribution that satisfies Zipf's law: the frequency of an item is inversely proportional to its rank in a frequency table. .. note:: New code should use the ``zipf`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- a : float or array_like of floats Distribution parameter. Must be greater than 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``a`` is a scalar. Otherwise, ``np.array(a).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Zipf distribution. See Also -------- scipy.stats.zipf : probability density function, distribution, or cumulative density function, etc. Generator.zipf: which should be used for new code. Notes ----- The probability density for the Zipf distribution is .. math:: p(x) = \frac{x^{-a}}{\zeta(a)}, where :math:`\zeta` is the Riemann Zeta function. It is named for the American linguist George Kingsley Zipf, who noted that the frequency of any word in a sample of a language is inversely proportional to its rank in the frequency table. References ---------- .. [1] Zipf, G. K., "Selected Studies of the Principle of Relative Frequency in Language," Cambridge, MA: Harvard Univ. Press, 1932. Examples -------- Draw samples from the distribution: >>> a = 2. # parameter >>> s = np.random.zipf(a, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> from scipy import special # doctest: +SKIP Truncate s values at 50 so plot is interesting: >>> count, bins, ignored = plt.hist(s[s<50], 50, density=True) >>> x = np.arange(1., 50.) >>> y = x**(-a) / special.zetac(a) # doctest: +SKIP >>> plt.plot(x, y/max(y), linewidth=2, color='r') # doctest: +SKIP >>> plt.show() zipfzerosyou are shuffling a 'x must be an integer or at least 1-dimensional weibull(a, size=None) Draw samples from a Weibull distribution. Draw samples from a 1-parameter Weibull distribution with the given shape parameter `a`. .. math:: X = (-ln(U))^{1/a} Here, U is drawn from the uniform distribution over (0,1]. The more common 2-parameter Weibull, including a scale parameter :math:`\lambda` is just :math:`X = \lambda(-ln(U))^{1/a}`. .. note:: New code should use the ``weibull`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- a : float or array_like of floats Shape parameter of the distribution. Must be nonnegative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``a`` is a scalar. Otherwise, ``np.array(a).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Weibull distribution. See Also -------- scipy.stats.weibull_max scipy.stats.weibull_min scipy.stats.genextreme gumbel Generator.weibull: which should be used for new code. Notes ----- The Weibull (or Type III asymptotic extreme value distribution for smallest values, SEV Type III, or Rosin-Rammler distribution) is one of a class of Generalized Extreme Value (GEV) distributions used in modeling extreme value problems. This class includes the Gumbel and Frechet distributions. The probability density for the Weibull distribution is .. math:: p(x) = \frac{a} {\lambda}(\frac{x}{\lambda})^{a-1}e^{-(x/\lambda)^a}, where :math:`a` is the shape and :math:`\lambda` the scale. The function has its peak (the mode) at :math:`\lambda(\frac{a-1}{a})^{1/a}`. When ``a = 1``, the Weibull distribution reduces to the exponential distribution. References ---------- .. [1] Waloddi Weibull, Royal Technical University, Stockholm, 1939 "A Statistical Theory Of The Strength Of Materials", Ingeniorsvetenskapsakademiens Handlingar Nr 151, 1939, Generalstabens Litografiska Anstalts Forlag, Stockholm. .. [2] Waloddi Weibull, "A Statistical Distribution Function of Wide Applicability", Journal Of Applied Mechanics ASME Paper 1951. .. [3] Wikipedia, "Weibull distribution", https://en.wikipedia.org/wiki/Weibull_distribution Examples -------- Draw samples from the distribution: >>> a = 5. # shape >>> s = np.random.weibull(a, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> x = np.arange(1,100.)/50. >>> def weib(x,n,a): ... return (a / n) * (x / n)**(a - 1) * np.exp(-(x / n)**a) >>> count, bins, ignored = plt.hist(np.random.weibull(5.,1000)) >>> x = np.arange(1,100.)/50. >>> scale = count.max()/weib(x, 1., 5.).max() >>> plt.plot(x, weib(x, 1., 5.)*scale) >>> plt.show() weibullwarningswarn wald(mean, scale, size=None) Draw samples from a Wald, or inverse Gaussian, distribution. As the scale approaches infinity, the distribution becomes more like a Gaussian. Some references claim that the Wald is an inverse Gaussian with mean equal to 1, but this is by no means universal. The inverse Gaussian distribution was first studied in relationship to Brownian motion. In 1956 M.C.K. Tweedie used the name inverse Gaussian because there is an inverse relationship between the time to cover a unit distance and distance covered in unit time. .. note:: New code should use the ``wald`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- mean : float or array_like of floats Distribution mean, must be > 0. scale : float or array_like of floats Scale parameter, must be > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``mean`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(mean, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Wald distribution. See Also -------- Generator.wald: which should be used for new code. Notes ----- The probability density function for the Wald distribution is .. math:: P(x;mean,scale) = \sqrt{\frac{scale}{2\pi x^3}}e^ \frac{-scale(x-mean)^2}{2\cdotp mean^2x} As noted above the inverse Gaussian distribution first arise from attempts to model Brownian motion. It is also a competitor to the Weibull for use in reliability modeling and modeling stock returns and interest rate processes. References ---------- .. [1] Brighton Webs Ltd., Wald Distribution, https://web.archive.org/web/20090423014010/http://www.brighton-webs.co.uk:80/distributions/wald.asp .. [2] Chhikara, Raj S., and Folks, J. Leroy, "The Inverse Gaussian Distribution: Theory : Methodology, and Applications", CRC Press, 1988. .. [3] Wikipedia, "Inverse Gaussian distribution" https://en.wikipedia.org/wiki/Inverse_Gaussian_distribution Examples -------- Draw values from the distribution and plot the histogram: >>> import matplotlib.pyplot as plt >>> h = plt.hist(np.random.wald(3, 2, 100000), bins=200, density=True) >>> plt.show() wald vonmises(mu, kappa, size=None) Draw samples from a von Mises distribution. Samples are drawn from a von Mises distribution with specified mode (mu) and dispersion (kappa), on the interval [-pi, pi]. The von Mises distribution (also known as the circular normal distribution) is a continuous probability distribution on the unit circle. It may be thought of as the circular analogue of the normal distribution. .. note:: New code should use the ``vonmises`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- mu : float or array_like of floats Mode ("center") of the distribution. kappa : float or array_like of floats Dispersion of the distribution, has to be >=0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``mu`` and ``kappa`` are both scalars. Otherwise, ``np.broadcast(mu, kappa).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized von Mises distribution. See Also -------- scipy.stats.vonmises : probability density function, distribution, or cumulative density function, etc. Generator.vonmises: which should be used for new code. Notes ----- The probability density for the von Mises distribution is .. math:: p(x) = \frac{e^{\kappa cos(x-\mu)}}{2\pi I_0(\kappa)}, where :math:`\mu` is the mode and :math:`\kappa` the dispersion, and :math:`I_0(\kappa)` is the modified Bessel function of order 0. The von Mises is named for Richard Edler von Mises, who was born in Austria-Hungary, in what is now the Ukraine. He fled to the United States in 1939 and became a professor at Harvard. He worked in probability theory, aerodynamics, fluid mechanics, and philosophy of science. References ---------- .. [1] Abramowitz, M. and Stegun, I. A. (Eds.). "Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing," New York: Dover, 1972. .. [2] von Mises, R., "Mathematical Theory of Probability and Statistics", New York: Academic Press, 1964. Examples -------- Draw samples from the distribution: >>> mu, kappa = 0.0, 4.0 # mean and dispersion >>> s = np.random.vonmises(mu, kappa, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> from scipy.special import i0 # doctest: +SKIP >>> plt.hist(s, 50, density=True) >>> x = np.linspace(-np.pi, np.pi, num=51) >>> y = np.exp(kappa*np.cos(x-mu))/(2*np.pi*i0(kappa)) # doctest: +SKIP >>> plt.plot(x, y, linewidth=2, color='r') # doctest: +SKIP >>> plt.show() vonmisesunsafeunique uniform(low=0.0, high=1.0, size=None) Draw samples from a uniform distribution. Samples are uniformly distributed over the half-open interval ``[low, high)`` (includes low, but excludes high). In other words, any value within the given interval is equally likely to be drawn by `uniform`. .. note:: New code should use the ``uniform`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- low : float or array_like of floats, optional Lower boundary of the output interval. All values generated will be greater than or equal to low. The default value is 0. high : float or array_like of floats Upper boundary of the output interval. All values generated will be less than or equal to high. The default value is 1.0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``low`` and ``high`` are both scalars. Otherwise, ``np.broadcast(low, high).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized uniform distribution. See Also -------- randint : Discrete uniform distribution, yielding integers. random_integers : Discrete uniform distribution over the closed interval ``[low, high]``. random_sample : Floats uniformly distributed over ``[0, 1)``. random : Alias for `random_sample`. rand : Convenience function that accepts dimensions as input, e.g., ``rand(2,2)`` would generate a 2-by-2 array of floats, uniformly distributed over ``[0, 1)``. Generator.uniform: which should be used for new code. Notes ----- The probability density function of the uniform distribution is .. math:: p(x) = \frac{1}{b - a} anywhere within the interval ``[a, b)``, and zero elsewhere. When ``high`` == ``low``, values of ``low`` will be returned. If ``high`` < ``low``, the results are officially undefined and may eventually raise an error, i.e. do not rely on this function to behave when passed arguments satisfying that inequality condition. The ``high`` limit may be included in the returned array of floats due to floating-point rounding in the equation ``low + (high-low) * random_sample()``. For example: >>> x = np.float32(5*0.99999999) >>> x 5.0 Examples -------- Draw samples from the distribution: >>> s = np.random.uniform(-1,0,1000) All values are within the given interval: >>> np.all(s >= -1) True >>> np.all(s < 0) True Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 15, density=True) >>> plt.plot(bins, np.ones_like(bins), linewidth=2, color='r') >>> plt.show() uniformuint8uint64uint32uint16>> import matplotlib.pyplot as plt >>> h = plt.hist(np.random.triangular(-3, 0, 8, 100000), bins=200, ... density=True) >>> plt.show() triangular tomaxint(size=None) Return a sample of uniformly distributed random integers in the interval [0, ``np.iinfo(np.int_).max``]. The `np.int_` type translates to the C long integer type and its precision is platform dependent. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray Drawn samples, with shape `size`. See Also -------- randint : Uniform sampling over a given half-open interval of integers. random_integers : Uniform sampling over a given closed interval of integers. Examples -------- >>> rs = np.random.RandomState() # need a RandomState object >>> rs.tomaxint((2,2,2)) array([[[1170048599, 1600360186], # random [ 739731006, 1947757578]], [[1871712945, 752307660], [1601631370, 1479324245]]]) >>> rs.tomaxint((2,2,2)) < np.iinfo(np.int_).max array([[[ True, True], [ True, True]], [[ True, True], [ True, True]]]) toltobytes__test__takesvdsum(pvals[:-1].astype(np.float64)) > 1.0. The pvals array is cast to 64-bit floating point prior to checking the sum. Precision changes when casting may cause problems even if the sum of the original pvals is valid.sum(pvals[:-1]) > 1.0sumsubtractstrides__str__state must be a dict or a tuple.state dictionary is not valid.state standard_t(df, size=None) Draw samples from a standard Student's t distribution with `df` degrees of freedom. A special case of the hyperbolic distribution. As `df` gets large, the result resembles that of the standard normal distribution (`standard_normal`). .. note:: New code should use the ``standard_t`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- df : float or array_like of floats Degrees of freedom, must be > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``df`` is a scalar. Otherwise, ``np.array(df).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized standard Student's t distribution. See Also -------- Generator.standard_t: which should be used for new code. Notes ----- The probability density function for the t distribution is .. math:: P(x, df) = \frac{\Gamma(\frac{df+1}{2})}{\sqrt{\pi df} \Gamma(\frac{df}{2})}\Bigl( 1+\frac{x^2}{df} \Bigr)^{-(df+1)/2} The t test is based on an assumption that the data come from a Normal distribution. The t test provides a way to test whether the sample mean (that is the mean calculated from the data) is a good estimate of the true mean. The derivation of the t-distribution was first published in 1908 by William Gosset while working for the Guinness Brewery in Dublin. Due to proprietary issues, he had to publish under a pseudonym, and so he used the name Student. References ---------- .. [1] Dalgaard, Peter, "Introductory Statistics With R", Springer, 2002. .. [2] Wikipedia, "Student's t-distribution" https://en.wikipedia.org/wiki/Student's_t-distribution Examples -------- From Dalgaard page 83 [1]_, suppose the daily energy intake for 11 women in kilojoules (kJ) is: >>> intake = np.array([5260., 5470, 5640, 6180, 6390, 6515, 6805, 7515, \ ... 7515, 8230, 8770]) Does their energy intake deviate systematically from the recommended value of 7725 kJ? Our null hypothesis will be the absence of deviation, and the alternate hypothesis will be the presence of an effect that could be either positive or negative, hence making our test 2-tailed. Because we are estimating the mean and we have N=11 values in our sample, we have N-1=10 degrees of freedom. We set our significance level to 95% and compute the t statistic using the empirical mean and empirical standard deviation of our intake. We use a ddof of 1 to base the computation of our empirical standard deviation on an unbiased estimate of the variance (note: the final estimate is not unbiased due to the concave nature of the square root). >>> np.mean(intake) 6753.636363636364 >>> intake.std(ddof=1) 1142.1232221373727 >>> t = (np.mean(intake)-7725)/(intake.std(ddof=1)/np.sqrt(len(intake))) >>> t -2.8207540608310198 We draw 1000000 samples from Student's t distribution with the adequate degrees of freedom. >>> import matplotlib.pyplot as plt >>> s = np.random.standard_t(10, size=1000000) >>> h = plt.hist(s, bins=100, density=True) Does our t statistic land in one of the two critical regions found at both tails of the distribution? >>> np.sum(np.abs(t) < np.abs(s)) / float(len(s)) 0.018318 #random < 0.05, statistic is in critical region The probability value for this 2-tailed test is about 1.83%, which is lower than the 5% pre-determined significance threshold. Therefore, the probability of observing values as extreme as our intake conditionally on the null hypothesis being true is too low, and we reject the null hypothesis of no deviation. standard_t standard_normal(size=None) Draw samples from a standard Normal distribution (mean=0, stdev=1). .. note:: New code should use the ``standard_normal`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray A floating-point array of shape ``size`` of drawn samples, or a single sample if ``size`` was not specified. See Also -------- normal : Equivalent function with additional ``loc`` and ``scale`` arguments for setting the mean and standard deviation. Generator.standard_normal: which should be used for new code. Notes ----- For random samples from :math:`N(\mu, \sigma^2)`, use one of:: mu + sigma * np.random.standard_normal(size=...) np.random.normal(mu, sigma, size=...) Examples -------- >>> np.random.standard_normal() 2.1923875335537315 #random >>> s = np.random.standard_normal(8000) >>> s array([ 0.6888893 , 0.78096262, -0.89086505, ..., 0.49876311, # random -0.38672696, -0.4685006 ]) # random >>> s.shape (8000,) >>> s = np.random.standard_normal(size=(3, 4, 2)) >>> s.shape (3, 4, 2) Two-by-four array of samples from :math:`N(3, 6.25)`: >>> 3 + 2.5 * np.random.standard_normal(size=(2, 4)) array([[-4.49401501, 4.00950034, -1.81814867, 7.29718677], # random [ 0.39924804, 4.68456316, 4.99394529, 4.84057254]]) # random standard_normal standard_gamma(shape, size=None) Draw samples from a standard Gamma distribution. Samples are drawn from a Gamma distribution with specified parameters, shape (sometimes designated "k") and scale=1. .. note:: New code should use the ``standard_gamma`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- shape : float or array_like of floats Parameter, must be non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``shape`` is a scalar. Otherwise, ``np.array(shape).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized standard gamma distribution. See Also -------- scipy.stats.gamma : probability density function, distribution or cumulative density function, etc. Generator.standard_gamma: which should be used for new code. Notes ----- The probability density for the Gamma distribution is .. math:: p(x) = x^{k-1}\frac{e^{-x/\theta}}{\theta^k\Gamma(k)}, where :math:`k` is the shape and :math:`\theta` the scale, and :math:`\Gamma` is the Gamma function. The Gamma distribution is often used to model the times to failure of electronic components, and arises naturally in processes for which the waiting times between Poisson distributed events are relevant. References ---------- .. [1] Weisstein, Eric W. "Gamma Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/GammaDistribution.html .. [2] Wikipedia, "Gamma distribution", https://en.wikipedia.org/wiki/Gamma_distribution Examples -------- Draw samples from the distribution: >>> shape, scale = 2., 1. # mean and width >>> s = np.random.standard_gamma(shape, 1000000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> import scipy.special as sps # doctest: +SKIP >>> count, bins, ignored = plt.hist(s, 50, density=True) >>> y = bins**(shape-1) * ((np.exp(-bins/scale))/ # doctest: +SKIP ... (sps.gamma(shape) * scale**shape)) >>> plt.plot(bins, y, linewidth=2, color='r') # doctest: +SKIP >>> plt.show() standard_gamma standard_exponential(size=None) Draw samples from the standard exponential distribution. `standard_exponential` is identical to the exponential distribution with a scale parameter of 1. .. note:: New code should use the ``standard_exponential`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray Drawn samples. See Also -------- Generator.standard_exponential: which should be used for new code. Examples -------- Output a 3x8000 array: >>> n = np.random.standard_exponential((3, 8000)) standard_exponential standard_cauchy(size=None) Draw samples from a standard Cauchy distribution with mode = 0. Also known as the Lorentz distribution. .. note:: New code should use the ``standard_cauchy`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The drawn samples. See Also -------- Generator.standard_cauchy: which should be used for new code. Notes ----- The probability density function for the full Cauchy distribution is .. math:: P(x; x_0, \gamma) = \frac{1}{\pi \gamma \bigl[ 1+ (\frac{x-x_0}{\gamma})^2 \bigr] } and the Standard Cauchy distribution just sets :math:`x_0=0` and :math:`\gamma=1` The Cauchy distribution arises in the solution to the driven harmonic oscillator problem, and also describes spectral line broadening. It also describes the distribution of values at which a line tilted at a random angle will cut the x axis. When studying hypothesis tests that assume normality, seeing how the tests perform on data from a Cauchy distribution is a good indicator of their sensitivity to a heavy-tailed distribution, since the Cauchy looks very much like a Gaussian distribution, but with heavier tails. References ---------- .. [1] NIST/SEMATECH e-Handbook of Statistical Methods, "Cauchy Distribution", https://www.itl.nist.gov/div898/handbook/eda/section3/eda3663.htm .. [2] Weisstein, Eric W. "Cauchy Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/CauchyDistribution.html .. [3] Wikipedia, "Cauchy distribution" https://en.wikipedia.org/wiki/Cauchy_distribution Examples -------- Draw samples and plot the distribution: >>> import matplotlib.pyplot as plt >>> s = np.random.standard_cauchy(1000000) >>> s = s[(s>-25) & (s<25)] # truncate distribution so it plots well >>> plt.hist(s, bins=100) >>> plt.show() standard_cauchystacklevelsqrtsortsizesigmaside shuffle(x) Modify a sequence in-place by shuffling its contents. This function only shuffles the array along the first axis of a multi-dimensional array. The order of sub-arrays is changed but their contents remains the same. .. note:: New code should use the ``shuffle`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- x : ndarray or MutableSequence The array, list or mutable sequence to be shuffled. Returns ------- None See Also -------- Generator.shuffle: which should be used for new code. Examples -------- >>> arr = np.arange(10) >>> np.random.shuffle(arr) >>> arr [1 7 5 2 9 4 3 6 0 8] # random Multi-dimensional arrays are only shuffled along the first axis: >>> arr = np.arange(9).reshape((3, 3)) >>> np.random.shuffle(arr) >>> arr array([[3, 4, 5], # random [6, 7, 8], [0, 1, 2]]) shuffleshapeset_state can only be used with legacy MT19937state instances.set_state seed(self, seed=None) Reseed a legacy MT19937 BitGenerator Notes ----- This is a convenience, legacy function. The best practice is to **not** reseed a BitGenerator, rather to recreate a new one. This method is here for legacy reasons. This example demonstrates best practice. >>> from numpy.random import MT19937 >>> from numpy.random import RandomState, SeedSequence >>> rs = RandomState(MT19937(SeedSequence(123456789))) # Later, you want to restart the stream >>> rs = RandomState(MT19937(SeedSequence(987654321))) seedsearchsortedscalesamplertolrightreversedreturn_indexreshapereplacereduce rayleigh(scale=1.0, size=None) Draw samples from a Rayleigh distribution. The :math:`\chi` and Weibull distributions are generalizations of the Rayleigh. .. note:: New code should use the ``rayleigh`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- scale : float or array_like of floats, optional Scale, also equals the mode. Must be non-negative. Default is 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``scale`` is a scalar. Otherwise, ``np.array(scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Rayleigh distribution. See Also -------- Generator.rayleigh: which should be used for new code. Notes ----- The probability density function for the Rayleigh distribution is .. math:: P(x;scale) = \frac{x}{scale^2}e^{\frac{-x^2}{2 \cdotp scale^2}} The Rayleigh distribution would arise, for example, if the East and North components of the wind velocity had identical zero-mean Gaussian distributions. Then the wind speed would have a Rayleigh distribution. References ---------- .. [1] Brighton Webs Ltd., "Rayleigh Distribution," https://web.archive.org/web/20090514091424/http://brighton-webs.co.uk:80/distributions/rayleigh.asp .. [2] Wikipedia, "Rayleigh distribution" https://en.wikipedia.org/wiki/Rayleigh_distribution Examples -------- Draw values from the distribution and plot the histogram >>> from matplotlib.pyplot import hist >>> values = hist(np.random.rayleigh(3, 100000), bins=200, density=True) Wave heights tend to follow a Rayleigh distribution. If the mean wave height is 1 meter, what fraction of waves are likely to be larger than 3 meters? >>> meanvalue = 1 >>> modevalue = np.sqrt(2 / np.pi) * meanvalue >>> s = np.random.rayleigh(modevalue, 1000000) The percentage of waves larger than 3 meters is: >>> 100.*sum(s>3)/1000000. 0.087300000000000003 # random rayleighravelrangeranf__randomstate_ctor random_sample(size=None) Return random floats in the half-open interval [0.0, 1.0). Results are from the "continuous uniform" distribution over the stated interval. To sample :math:`Unif[a, b), b > a` multiply the output of `random_sample` by `(b-a)` and add `a`:: (b - a) * random_sample() + a .. note:: New code should use the ``random`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray of floats Array of random floats of shape `size` (unless ``size=None``, in which case a single float is returned). See Also -------- Generator.random: which should be used for new code. Examples -------- >>> np.random.random_sample() 0.47108547995356098 # random >>> type(np.random.random_sample()) >>> np.random.random_sample((5,)) array([ 0.30220482, 0.86820401, 0.1654503 , 0.11659149, 0.54323428]) # random Three-by-two array of random numbers from [-5, 0): >>> 5 * np.random.random_sample((3, 2)) - 5 array([[-3.99149989, -0.52338984], # random [-2.99091858, -0.79479508], [-1.23204345, -1.75224494]]) random_sample random_integers(low, high=None, size=None) Random integers of type `np.int_` between `low` and `high`, inclusive. Return random integers of type `np.int_` from the "discrete uniform" distribution in the closed interval [`low`, `high`]. If `high` is None (the default), then results are from [1, `low`]. The `np.int_` type translates to the C long integer type and its precision is platform dependent. This function has been deprecated. Use randint instead. .. deprecated:: 1.11.0 Parameters ---------- low : int Lowest (signed) integer to be drawn from the distribution (unless ``high=None``, in which case this parameter is the *highest* such integer). high : int, optional If provided, the largest (signed) integer to be drawn from the distribution (see above for behavior if ``high=None``). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : int or ndarray of ints `size`-shaped array of random integers from the appropriate distribution, or a single such random int if `size` not provided. See Also -------- randint : Similar to `random_integers`, only for the half-open interval [`low`, `high`), and 0 is the lowest value if `high` is omitted. Notes ----- To sample from N evenly spaced floating-point numbers between a and b, use:: a + (b - a) * (np.random.random_integers(N) - 1) / (N - 1.) Examples -------- >>> np.random.random_integers(5) 4 # random >>> type(np.random.random_integers(5)) >>> np.random.random_integers(5, size=(3,2)) array([[5, 4], # random [3, 3], [4, 5]]) Choose five random numbers from the set of five evenly-spaced numbers between 0 and 2.5, inclusive (*i.e.*, from the set :math:`{0, 5/8, 10/8, 15/8, 20/8}`): >>> 2.5 * (np.random.random_integers(5, size=(5,)) - 1) / 4. array([ 0.625, 1.25 , 0.625, 0.625, 2.5 ]) # random Roll two six sided dice 1000 times and sum the results: >>> d1 = np.random.random_integers(1, 6, 1000) >>> d2 = np.random.random_integers(1, 6, 1000) >>> dsums = d1 + d2 Display results as a histogram: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(dsums, 11, density=True) >>> plt.show() random_integersrandom randn(d0, d1, ..., dn) Return a sample (or samples) from the "standard normal" distribution. .. note:: This is a convenience function for users porting code from Matlab, and wraps `standard_normal`. That function takes a tuple to specify the size of the output, which is consistent with other NumPy functions like `numpy.zeros` and `numpy.ones`. .. note:: New code should use the ``standard_normal`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. If positive int_like arguments are provided, `randn` generates an array of shape ``(d0, d1, ..., dn)``, filled with random floats sampled from a univariate "normal" (Gaussian) distribution of mean 0 and variance 1. A single float randomly sampled from the distribution is returned if no argument is provided. Parameters ---------- d0, d1, ..., dn : int, optional The dimensions of the returned array, must be non-negative. If no argument is given a single Python float is returned. Returns ------- Z : ndarray or float A ``(d0, d1, ..., dn)``-shaped array of floating-point samples from the standard normal distribution, or a single such float if no parameters were supplied. See Also -------- standard_normal : Similar, but takes a tuple as its argument. normal : Also accepts mu and sigma arguments. Generator.standard_normal: which should be used for new code. Notes ----- For random samples from :math:`N(\mu, \sigma^2)`, use: ``sigma * np.random.randn(...) + mu`` Examples -------- >>> np.random.randn() 2.1923875335537315 # random Two-by-four array of samples from N(3, 6.25): >>> 3 + 2.5 * np.random.randn(2, 4) array([[-4.49401501, 4.00950034, -1.81814867, 7.29718677], # random [ 0.39924804, 4.68456316, 4.99394529, 4.84057254]]) # random randn randint(low, high=None, size=None, dtype=int) Return random integers from `low` (inclusive) to `high` (exclusive). Return random integers from the "discrete uniform" distribution of the specified dtype in the "half-open" interval [`low`, `high`). If `high` is None (the default), then results are from [0, `low`). .. note:: New code should use the ``integers`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- low : int or array-like of ints Lowest (signed) integers to be drawn from the distribution (unless ``high=None``, in which case this parameter is one above the *highest* such integer). high : int or array-like of ints, optional If provided, one above the largest (signed) integer to be drawn from the distribution (see above for behavior if ``high=None``). If array-like, must contain integer values size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. dtype : dtype, optional Desired dtype of the result. Byteorder must be native. The default value is int. .. versionadded:: 1.11.0 Returns ------- out : int or ndarray of ints `size`-shaped array of random integers from the appropriate distribution, or a single such random int if `size` not provided. See Also -------- random_integers : similar to `randint`, only for the closed interval [`low`, `high`], and 1 is the lowest value if `high` is omitted. Generator.integers: which should be used for new code. Examples -------- >>> np.random.randint(2, size=10) array([1, 0, 0, 0, 1, 1, 0, 0, 1, 0]) # random >>> np.random.randint(1, size=10) array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0]) Generate a 2 x 4 array of ints between 0 and 4, inclusive: >>> np.random.randint(5, size=(2, 4)) array([[4, 0, 2, 1], # random [3, 2, 2, 0]]) Generate a 1 x 3 array with 3 different upper bounds >>> np.random.randint(1, [3, 5, 10]) array([2, 2, 9]) # random Generate a 1 by 3 array with 3 different lower bounds >>> np.random.randint([1, 5, 7], 10) array([9, 8, 7]) # random Generate a 2 by 4 array using broadcasting with dtype of uint8 >>> np.random.randint([1, 3, 5, 7], [[10], [20]], dtype=np.uint8) array([[ 8, 6, 9, 7], # random [ 1, 16, 9, 12]], dtype=uint8) randint rand(d0, d1, ..., dn) Random values in a given shape. .. note:: This is a convenience function for users porting code from Matlab, and wraps `random_sample`. That function takes a tuple to specify the size of the output, which is consistent with other NumPy functions like `numpy.zeros` and `numpy.ones`. Create an array of the given shape and populate it with random samples from a uniform distribution over ``[0, 1)``. Parameters ---------- d0, d1, ..., dn : int, optional The dimensions of the returned array, must be non-negative. If no argument is given a single Python float is returned. Returns ------- out : ndarray, shape ``(d0, d1, ..., dn)`` Random values. See Also -------- random Examples -------- >>> np.random.rand(3,2) array([[ 0.14022471, 0.96360618], #random [ 0.37601032, 0.25528411], #random [ 0.49313049, 0.94909878]]) #random _randrandraise__pyx_vtable__pvalsprodprobabilities do not sum to 1probabilities contain NaNprobabilities are not non-negative power(a, size=None) Draws samples in [0, 1] from a power distribution with positive exponent a - 1. Also known as the power function distribution. .. note:: New code should use the ``power`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- a : float or array_like of floats Parameter of the distribution. Must be non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``a`` is a scalar. Otherwise, ``np.array(a).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized power distribution. Raises ------ ValueError If a < 1. See Also -------- Generator.power: which should be used for new code. Notes ----- The probability density function is .. math:: P(x; a) = ax^{a-1}, 0 \le x \le 1, a>0. The power function distribution is just the inverse of the Pareto distribution. It may also be seen as a special case of the Beta distribution. It is used, for example, in modeling the over-reporting of insurance claims. References ---------- .. [1] Christian Kleiber, Samuel Kotz, "Statistical size distributions in economics and actuarial sciences", Wiley, 2003. .. [2] Heckert, N. A. and Filliben, James J. "NIST Handbook 148: Dataplot Reference Manual, Volume 2: Let Subcommands and Library Functions", National Institute of Standards and Technology Handbook Series, June 2003. https://www.itl.nist.gov/div898/software/dataplot/refman2/auxillar/powpdf.pdf Examples -------- Draw samples from the distribution: >>> a = 5. # shape >>> samples = 1000 >>> s = np.random.power(a, samples) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, bins=30) >>> x = np.linspace(0, 1, 100) >>> y = a*x**(a-1.) >>> normed_y = samples*np.diff(bins)[0]*y >>> plt.plot(x, normed_y) >>> plt.show() Compare the power function distribution to the inverse of the Pareto. >>> from scipy import stats # doctest: +SKIP >>> rvs = np.random.power(5, 1000000) >>> rvsp = np.random.pareto(5, 1000000) >>> xx = np.linspace(0,1,100) >>> powpdf = stats.powerlaw.pdf(xx,5) # doctest: +SKIP >>> plt.figure() >>> plt.hist(rvs, bins=50, density=True) >>> plt.plot(xx,powpdf,'r-') # doctest: +SKIP >>> plt.title('np.random.power(5)') >>> plt.figure() >>> plt.hist(1./(1.+rvsp), bins=50, density=True) >>> plt.plot(xx,powpdf,'r-') # doctest: +SKIP >>> plt.title('inverse of 1 + np.random.pareto(5)') >>> plt.figure() >>> plt.hist(1./(1.+rvsp), bins=50, density=True) >>> plt.plot(xx,powpdf,'r-') # doctest: +SKIP >>> plt.title('inverse of stats.pareto(5)') powerpos_poisson_lam_max poisson(lam=1.0, size=None) Draw samples from a Poisson distribution. The Poisson distribution is the limit of the binomial distribution for large N. .. note:: New code should use the ``poisson`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- lam : float or array_like of floats Expected number of events occurring in a fixed-time interval, must be >= 0. A sequence must be broadcastable over the requested size. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``lam`` is a scalar. Otherwise, ``np.array(lam).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Poisson distribution. See Also -------- Generator.poisson: which should be used for new code. Notes ----- The Poisson distribution .. math:: f(k; \lambda)=\frac{\lambda^k e^{-\lambda}}{k!} For events with an expected separation :math:`\lambda` the Poisson distribution :math:`f(k; \lambda)` describes the probability of :math:`k` events occurring within the observed interval :math:`\lambda`. Because the output is limited to the range of the C int64 type, a ValueError is raised when `lam` is within 10 sigma of the maximum representable value. References ---------- .. [1] Weisstein, Eric W. "Poisson Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/PoissonDistribution.html .. [2] Wikipedia, "Poisson distribution", https://en.wikipedia.org/wiki/Poisson_distribution Examples -------- Draw samples from the distribution: >>> import numpy as np >>> s = np.random.poisson(5, 10000) Display histogram of the sample: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 14, density=True) >>> plt.show() Draw each 100 values for lambda 100 and 500: >>> s = np.random.poisson(lam=(100., 500.), size=(100, 2)) poisson_pickle permutation(x) Randomly permute a sequence, or return a permuted range. If `x` is a multi-dimensional array, it is only shuffled along its first index. .. note:: New code should use the ``permutation`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- x : int or array_like If `x` is an integer, randomly permute ``np.arange(x)``. If `x` is an array, make a copy and shuffle the elements randomly. Returns ------- out : ndarray Permuted sequence or array range. See Also -------- Generator.permutation: which should be used for new code. Examples -------- >>> np.random.permutation(10) array([1, 7, 4, 3, 0, 9, 2, 5, 8, 6]) # random >>> np.random.permutation([1, 4, 9, 12, 15]) array([15, 1, 9, 4, 12]) # random >>> arr = np.arange(9).reshape((3, 3)) >>> np.random.permutation(arr) array([[6, 7, 8], # random [0, 1, 2], [3, 4, 5]]) permutation pareto(a, size=None) Draw samples from a Pareto II or Lomax distribution with specified shape. The Lomax or Pareto II distribution is a shifted Pareto distribution. The classical Pareto distribution can be obtained from the Lomax distribution by adding 1 and multiplying by the scale parameter ``m`` (see Notes). The smallest value of the Lomax distribution is zero while for the classical Pareto distribution it is ``mu``, where the standard Pareto distribution has location ``mu = 1``. Lomax can also be considered as a simplified version of the Generalized Pareto distribution (available in SciPy), with the scale set to one and the location set to zero. The Pareto distribution must be greater than zero, and is unbounded above. It is also known as the "80-20 rule". In this distribution, 80 percent of the weights are in the lowest 20 percent of the range, while the other 20 percent fill the remaining 80 percent of the range. .. note:: New code should use the ``pareto`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- a : float or array_like of floats Shape of the distribution. Must be positive. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``a`` is a scalar. Otherwise, ``np.array(a).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Pareto distribution. See Also -------- scipy.stats.lomax : probability density function, distribution or cumulative density function, etc. scipy.stats.genpareto : probability density function, distribution or cumulative density function, etc. Generator.pareto: which should be used for new code. Notes ----- The probability density for the Pareto distribution is .. math:: p(x) = \frac{am^a}{x^{a+1}} where :math:`a` is the shape and :math:`m` the scale. The Pareto distribution, named after the Italian economist Vilfredo Pareto, is a power law probability distribution useful in many real world problems. Outside the field of economics it is generally referred to as the Bradford distribution. Pareto developed the distribution to describe the distribution of wealth in an economy. It has also found use in insurance, web page access statistics, oil field sizes, and many other problems, including the download frequency for projects in Sourceforge [1]_. It is one of the so-called "fat-tailed" distributions. References ---------- .. [1] Francis Hunt and Paul Johnson, On the Pareto Distribution of Sourceforge projects. .. [2] Pareto, V. (1896). Course of Political Economy. Lausanne. .. [3] Reiss, R.D., Thomas, M.(2001), Statistical Analysis of Extreme Values, Birkhauser Verlag, Basel, pp 23-30. .. [4] Wikipedia, "Pareto distribution", https://en.wikipedia.org/wiki/Pareto_distribution Examples -------- Draw samples from the distribution: >>> a, m = 3., 2. # shape and mode >>> s = (np.random.pareto(a, 1000) + 1) * m Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, _ = plt.hist(s, 100, density=True) >>> fit = a*m**a / bins**(a+1) >>> plt.plot(bins, max(count)*fit/max(fit), linewidth=2, color='r') >>> plt.show() pareto'p' must be 1-dimensionalpoperator' object which is not a subclass of 'Sequence'; `shuffle` is not guaranteed to behave correctly. E.g., non-numpy array/tensor objects with view semantics may contain duplicates after shuffling.object_numpy.random.mtrandnumpy.linalgnumpy.core.umath failed to importnumpy.core.multiarray failed to importnumpynsamplenp normal(loc=0.0, scale=1.0, size=None) Draw random samples from a normal (Gaussian) distribution. The probability density function of the normal distribution, first derived by De Moivre and 200 years later by both Gauss and Laplace independently [2]_, is often called the bell curve because of its characteristic shape (see the example below). The normal distributions occurs often in nature. For example, it describes the commonly occurring distribution of samples influenced by a large number of tiny, random disturbances, each with its own unique distribution [2]_. .. note:: New code should use the ``normal`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- loc : float or array_like of floats Mean ("centre") of the distribution. scale : float or array_like of floats Standard deviation (spread or "width") of the distribution. Must be non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``loc`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(loc, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized normal distribution. See Also -------- scipy.stats.norm : probability density function, distribution or cumulative density function, etc. Generator.normal: which should be used for new code. Notes ----- The probability density for the Gaussian distribution is .. math:: p(x) = \frac{1}{\sqrt{ 2 \pi \sigma^2 }} e^{ - \frac{ (x - \mu)^2 } {2 \sigma^2} }, where :math:`\mu` is the mean and :math:`\sigma` the standard deviation. The square of the standard deviation, :math:`\sigma^2`, is called the variance. The function has its peak at the mean, and its "spread" increases with the standard deviation (the function reaches 0.607 times its maximum at :math:`x + \sigma` and :math:`x - \sigma` [2]_). This implies that normal is more likely to return samples lying close to the mean, rather than those far away. References ---------- .. [1] Wikipedia, "Normal distribution", https://en.wikipedia.org/wiki/Normal_distribution .. [2] P. R. Peebles Jr., "Central Limit Theorem" in "Probability, Random Variables and Random Signal Principles", 4th ed., 2001, pp. 51, 51, 125. Examples -------- Draw samples from the distribution: >>> mu, sigma = 0, 0.1 # mean and standard deviation >>> s = np.random.normal(mu, sigma, 1000) Verify the mean and the variance: >>> abs(mu - np.mean(s)) 0.0 # may vary >>> abs(sigma - np.std(s, ddof=1)) 0.1 # may vary Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, density=True) >>> plt.plot(bins, 1/(sigma * np.sqrt(2 * np.pi)) * ... np.exp( - (bins - mu)**2 / (2 * sigma**2) ), ... linewidth=2, color='r') >>> plt.show() Two-by-four array of samples from N(3, 6.25): >>> np.random.normal(3, 2.5, size=(2, 4)) array([[-4.49401501, 4.00950034, -1.81814867, 7.29718677], # random [ 0.39924804, 4.68456316, 4.99394529, 4.84057254]]) # random normal noncentral_f(dfnum, dfden, nonc, size=None) Draw samples from the noncentral F distribution. Samples are drawn from an F distribution with specified parameters, `dfnum` (degrees of freedom in numerator) and `dfden` (degrees of freedom in denominator), where both parameters > 1. `nonc` is the non-centrality parameter. .. note:: New code should use the ``noncentral_f`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- dfnum : float or array_like of floats Numerator degrees of freedom, must be > 0. .. versionchanged:: 1.14.0 Earlier NumPy versions required dfnum > 1. dfden : float or array_like of floats Denominator degrees of freedom, must be > 0. nonc : float or array_like of floats Non-centrality parameter, the sum of the squares of the numerator means, must be >= 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``dfnum``, ``dfden``, and ``nonc`` are all scalars. Otherwise, ``np.broadcast(dfnum, dfden, nonc).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized noncentral Fisher distribution. See Also -------- Generator.noncentral_f: which should be used for new code. Notes ----- When calculating the power of an experiment (power = probability of rejecting the null hypothesis when a specific alternative is true) the non-central F statistic becomes important. When the null hypothesis is true, the F statistic follows a central F distribution. When the null hypothesis is not true, then it follows a non-central F statistic. References ---------- .. [1] Weisstein, Eric W. "Noncentral F-Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/NoncentralF-Distribution.html .. [2] Wikipedia, "Noncentral F-distribution", https://en.wikipedia.org/wiki/Noncentral_F-distribution Examples -------- In a study, testing for a specific alternative to the null hypothesis requires use of the Noncentral F distribution. We need to calculate the area in the tail of the distribution that exceeds the value of the F distribution for the null hypothesis. We'll plot the two probability distributions for comparison. >>> dfnum = 3 # between group deg of freedom >>> dfden = 20 # within groups degrees of freedom >>> nonc = 3.0 >>> nc_vals = np.random.noncentral_f(dfnum, dfden, nonc, 1000000) >>> NF = np.histogram(nc_vals, bins=50, density=True) >>> c_vals = np.random.f(dfnum, dfden, 1000000) >>> F = np.histogram(c_vals, bins=50, density=True) >>> import matplotlib.pyplot as plt >>> plt.plot(F[1][1:], F[0]) >>> plt.plot(NF[1][1:], NF[0]) >>> plt.show() noncentral_f noncentral_chisquare(df, nonc, size=None) Draw samples from a noncentral chi-square distribution. The noncentral :math:`\chi^2` distribution is a generalization of the :math:`\chi^2` distribution. .. note:: New code should use the ``noncentral_chisquare`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- df : float or array_like of floats Degrees of freedom, must be > 0. .. versionchanged:: 1.10.0 Earlier NumPy versions required dfnum > 1. nonc : float or array_like of floats Non-centrality, must be non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``df`` and ``nonc`` are both scalars. Otherwise, ``np.broadcast(df, nonc).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized noncentral chi-square distribution. See Also -------- Generator.noncentral_chisquare: which should be used for new code. Notes ----- The probability density function for the noncentral Chi-square distribution is .. math:: P(x;df,nonc) = \sum^{\infty}_{i=0} \frac{e^{-nonc/2}(nonc/2)^{i}}{i!} P_{Y_{df+2i}}(x), where :math:`Y_{q}` is the Chi-square with q degrees of freedom. References ---------- .. [1] Wikipedia, "Noncentral chi-squared distribution" https://en.wikipedia.org/wiki/Noncentral_chi-squared_distribution Examples -------- Draw values from the distribution and plot the histogram >>> import matplotlib.pyplot as plt >>> values = plt.hist(np.random.noncentral_chisquare(3, 20, 100000), ... bins=200, density=True) >>> plt.show() Draw values from a noncentral chisquare with very small noncentrality, and compare to a chisquare. >>> plt.figure() >>> values = plt.hist(np.random.noncentral_chisquare(3, .0000001, 100000), ... bins=np.arange(0., 25, .1), density=True) >>> values2 = plt.hist(np.random.chisquare(3, 100000), ... bins=np.arange(0., 25, .1), density=True) >>> plt.plot(values[1][0:-1], values[0]-values2[0], 'ob') >>> plt.show() Demonstrate how large values of non-centrality lead to a more symmetric distribution. >>> plt.figure() >>> values = plt.hist(np.random.noncentral_chisquare(3, 20, 100000), ... bins=200, density=True) >>> plt.show() noncentral_chisquarenoncngood + nbad < nsamplengoodnewbyteorder negative_binomial(n, p, size=None) Draw samples from a negative binomial distribution. Samples are drawn from a negative binomial distribution with specified parameters, `n` successes and `p` probability of success where `n` is > 0 and `p` is in the interval [0, 1]. .. note:: New code should use the ``negative_binomial`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- n : float or array_like of floats Parameter of the distribution, > 0. p : float or array_like of floats Parameter of the distribution, >= 0 and <=1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``n`` and ``p`` are both scalars. Otherwise, ``np.broadcast(n, p).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized negative binomial distribution, where each sample is equal to N, the number of failures that occurred before a total of n successes was reached. See Also -------- Generator.negative_binomial: which should be used for new code. Notes ----- The probability mass function of the negative binomial distribution is .. math:: P(N;n,p) = \frac{\Gamma(N+n)}{N!\Gamma(n)}p^{n}(1-p)^{N}, where :math:`n` is the number of successes, :math:`p` is the probability of success, :math:`N+n` is the number of trials, and :math:`\Gamma` is the gamma function. When :math:`n` is an integer, :math:`\frac{\Gamma(N+n)}{N!\Gamma(n)} = \binom{N+n-1}{N}`, which is the more common form of this term in the the pmf. The negative binomial distribution gives the probability of N failures given n successes, with a success on the last trial. If one throws a die repeatedly until the third time a "1" appears, then the probability distribution of the number of non-"1"s that appear before the third "1" is a negative binomial distribution. References ---------- .. [1] Weisstein, Eric W. "Negative Binomial Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/NegativeBinomialDistribution.html .. [2] Wikipedia, "Negative binomial distribution", https://en.wikipedia.org/wiki/Negative_binomial_distribution Examples -------- Draw samples from the distribution: A real world example. A company drills wild-cat oil exploration wells, each with an estimated probability of success of 0.1. What is the probability of having one success for each successive well, that is what is the probability of a single success after drilling 5 wells, after 6 wells, etc.? >>> s = np.random.negative_binomial(1, 0.1, 100000) >>> for i in range(1, 11): # doctest: +SKIP ... probability = sum(s>> mean = [0, 0] >>> cov = [[1, 0], [0, 100]] # diagonal covariance Diagonal covariance means that points are oriented along x or y-axis: >>> import matplotlib.pyplot as plt >>> x, y = np.random.multivariate_normal(mean, cov, 5000).T >>> plt.plot(x, y, 'x') >>> plt.axis('equal') >>> plt.show() Note that the covariance matrix must be positive semidefinite (a.k.a. nonnegative-definite). Otherwise, the behavior of this method is undefined and backwards compatibility is not guaranteed. References ---------- .. [1] Papoulis, A., "Probability, Random Variables, and Stochastic Processes," 3rd ed., New York: McGraw-Hill, 1991. .. [2] Duda, R. O., Hart, P. E., and Stork, D. G., "Pattern Classification," 2nd ed., New York: Wiley, 2001. Examples -------- >>> mean = (1, 2) >>> cov = [[1, 0], [0, 1]] >>> x = np.random.multivariate_normal(mean, cov, (3, 3)) >>> x.shape (3, 3, 2) The following is probably true, given that 0.6 is roughly twice the standard deviation: >>> list((x[0,0,:] - mean) < 0.6) [True, True] # random multivariate_normal multinomial(n, pvals, size=None) Draw samples from a multinomial distribution. The multinomial distribution is a multivariate generalization of the binomial distribution. Take an experiment with one of ``p`` possible outcomes. An example of such an experiment is throwing a dice, where the outcome can be 1 through 6. Each sample drawn from the distribution represents `n` such experiments. Its values, ``X_i = [X_0, X_1, ..., X_p]``, represent the number of times the outcome was ``i``. .. note:: New code should use the ``multinomial`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- n : int Number of experiments. pvals : sequence of floats, length p Probabilities of each of the ``p`` different outcomes. These must sum to 1 (however, the last element is always assumed to account for the remaining probability, as long as ``sum(pvals[:-1]) <= 1)``. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray The drawn samples, of shape *size*, if that was provided. If not, the shape is ``(N,)``. In other words, each entry ``out[i,j,...,:]`` is an N-dimensional value drawn from the distribution. See Also -------- Generator.multinomial: which should be used for new code. Examples -------- Throw a dice 20 times: >>> np.random.multinomial(20, [1/6.]*6, size=1) array([[4, 1, 7, 5, 2, 1]]) # random It landed 4 times on 1, once on 2, etc. Now, throw the dice 20 times, and 20 times again: >>> np.random.multinomial(20, [1/6.]*6, size=2) array([[3, 4, 3, 3, 4, 3], # random [2, 4, 3, 4, 0, 7]]) For the first run, we threw 3 times 1, 4 times 2, etc. For the second, we threw 2 times 1, 4 times 2, etc. A loaded die is more likely to land on number 6: >>> np.random.multinomial(100, [1/7.]*5 + [2/7.]) array([11, 16, 14, 17, 16, 26]) # random The probability inputs should be normalized. As an implementation detail, the value of the last entry is ignored and assumed to take up any leftover probability mass, but this should not be relied on. A biased coin which has twice as much weight on one side as on the other should be sampled like so: >>> np.random.multinomial(100, [1.0 / 3, 2.0 / 3]) # RIGHT array([38, 62]) # random not like: >>> np.random.multinomial(100, [1.0, 2.0]) # WRONG Traceback (most recent call last): ValueError: pvals < 0, pvals > 1 or pvals contains NaNs multinomialmumtrand.pyx_mt19937mode > rightmodemean must be 1 dimensionalmean and cov must have same lengthmeanmay_share_memory__main__lowlong logseries(p, size=None) Draw samples from a logarithmic series distribution. Samples are drawn from a log series distribution with specified shape parameter, 0 < ``p`` < 1. .. note:: New code should use the ``logseries`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- p : float or array_like of floats Shape parameter for the distribution. Must be in the range (0, 1). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``p`` is a scalar. Otherwise, ``np.array(p).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized logarithmic series distribution. See Also -------- scipy.stats.logser : probability density function, distribution or cumulative density function, etc. Generator.logseries: which should be used for new code. Notes ----- The probability density for the Log Series distribution is .. math:: P(k) = \frac{-p^k}{k \ln(1-p)}, where p = probability. The log series distribution is frequently used to represent species richness and occurrence, first proposed by Fisher, Corbet, and Williams in 1943 [2]. It may also be used to model the numbers of occupants seen in cars [3]. References ---------- .. [1] Buzas, Martin A.; Culver, Stephen J., Understanding regional species diversity through the log series distribution of occurrences: BIODIVERSITY RESEARCH Diversity & Distributions, Volume 5, Number 5, September 1999 , pp. 187-195(9). .. [2] Fisher, R.A,, A.S. Corbet, and C.B. Williams. 1943. The relation between the number of species and the number of individuals in a random sample of an animal population. Journal of Animal Ecology, 12:42-58. .. [3] D. J. Hand, F. Daly, D. Lunn, E. Ostrowski, A Handbook of Small Data Sets, CRC Press, 1994. .. [4] Wikipedia, "Logarithmic distribution", https://en.wikipedia.org/wiki/Logarithmic_distribution Examples -------- Draw samples from the distribution: >>> a = .6 >>> s = np.random.logseries(a, 10000) >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s) # plot against distribution >>> def logseries(k, p): ... return -p**k/(k*np.log(1-p)) >>> plt.plot(bins, logseries(bins, a)*count.max()/ ... logseries(bins, a).max(), 'r') >>> plt.show() logseries lognormal(mean=0.0, sigma=1.0, size=None) Draw samples from a log-normal distribution. Draw samples from a log-normal distribution with specified mean, standard deviation, and array shape. Note that the mean and standard deviation are not the values for the distribution itself, but of the underlying normal distribution it is derived from. .. note:: New code should use the ``lognormal`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- mean : float or array_like of floats, optional Mean value of the underlying normal distribution. Default is 0. sigma : float or array_like of floats, optional Standard deviation of the underlying normal distribution. Must be non-negative. Default is 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``mean`` and ``sigma`` are both scalars. Otherwise, ``np.broadcast(mean, sigma).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized log-normal distribution. See Also -------- scipy.stats.lognorm : probability density function, distribution, cumulative density function, etc. Generator.lognormal: which should be used for new code. Notes ----- A variable `x` has a log-normal distribution if `log(x)` is normally distributed. The probability density function for the log-normal distribution is: .. math:: p(x) = \frac{1}{\sigma x \sqrt{2\pi}} e^{(-\frac{(ln(x)-\mu)^2}{2\sigma^2})} where :math:`\mu` is the mean and :math:`\sigma` is the standard deviation of the normally distributed logarithm of the variable. A log-normal distribution results if a random variable is the *product* of a large number of independent, identically-distributed variables in the same way that a normal distribution results if the variable is the *sum* of a large number of independent, identically-distributed variables. References ---------- .. [1] Limpert, E., Stahel, W. A., and Abbt, M., "Log-normal Distributions across the Sciences: Keys and Clues," BioScience, Vol. 51, No. 5, May, 2001. https://stat.ethz.ch/~stahel/lognormal/bioscience.pdf .. [2] Reiss, R.D. and Thomas, M., "Statistical Analysis of Extreme Values," Basel: Birkhauser Verlag, 2001, pp. 31-32. Examples -------- Draw samples from the distribution: >>> mu, sigma = 3., 1. # mean and standard deviation >>> s = np.random.lognormal(mu, sigma, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 100, density=True, align='mid') >>> x = np.linspace(min(bins), max(bins), 10000) >>> pdf = (np.exp(-(np.log(x) - mu)**2 / (2 * sigma**2)) ... / (x * sigma * np.sqrt(2 * np.pi))) >>> plt.plot(x, pdf, linewidth=2, color='r') >>> plt.axis('tight') >>> plt.show() Demonstrate that taking the products of random samples from a uniform distribution can be fit well by a log-normal probability density function. >>> # Generate a thousand samples: each is the product of 100 random >>> # values, drawn from a normal distribution. >>> b = [] >>> for i in range(1000): ... a = 10. + np.random.standard_normal(100) ... b.append(np.product(a)) >>> b = np.array(b) / np.min(b) # scale values to be positive >>> count, bins, ignored = plt.hist(b, 100, density=True, align='mid') >>> sigma = np.std(np.log(b)) >>> mu = np.mean(np.log(b)) >>> x = np.linspace(min(bins), max(bins), 10000) >>> pdf = (np.exp(-(np.log(x) - mu)**2 / (2 * sigma**2)) ... / (x * sigma * np.sqrt(2 * np.pi))) >>> plt.plot(x, pdf, color='r', linewidth=2) >>> plt.show() lognormal logistic(loc=0.0, scale=1.0, size=None) Draw samples from a logistic distribution. Samples are drawn from a logistic distribution with specified parameters, loc (location or mean, also median), and scale (>0). .. note:: New code should use the ``logistic`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- loc : float or array_like of floats, optional Parameter of the distribution. Default is 0. scale : float or array_like of floats, optional Parameter of the distribution. Must be non-negative. Default is 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``loc`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(loc, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized logistic distribution. See Also -------- scipy.stats.logistic : probability density function, distribution or cumulative density function, etc. Generator.logistic: which should be used for new code. Notes ----- The probability density for the Logistic distribution is .. math:: P(x) = P(x) = \frac{e^{-(x-\mu)/s}}{s(1+e^{-(x-\mu)/s})^2}, where :math:`\mu` = location and :math:`s` = scale. The Logistic distribution is used in Extreme Value problems where it can act as a mixture of Gumbel distributions, in Epidemiology, and by the World Chess Federation (FIDE) where it is used in the Elo ranking system, assuming the performance of each player is a logistically distributed random variable. References ---------- .. [1] Reiss, R.-D. and Thomas M. (2001), "Statistical Analysis of Extreme Values, from Insurance, Finance, Hydrology and Other Fields," Birkhauser Verlag, Basel, pp 132-133. .. [2] Weisstein, Eric W. "Logistic Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/LogisticDistribution.html .. [3] Wikipedia, "Logistic-distribution", https://en.wikipedia.org/wiki/Logistic_distribution Examples -------- Draw samples from the distribution: >>> loc, scale = 10, 1 >>> s = np.random.logistic(loc, scale, 10000) >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, bins=50) # plot against distribution >>> def logist(x, loc, scale): ... return np.exp((loc-x)/scale)/(scale*(1+np.exp((loc-x)/scale))**2) >>> lgst_val = logist(bins, loc, scale) >>> plt.plot(bins, lgst_val * count.max() / lgst_val.max()) >>> plt.show() logisticlogical_orlocklocless_equalless_legacy_seedinglegacyleft == rightleft > modeleft laplace(loc=0.0, scale=1.0, size=None) Draw samples from the Laplace or double exponential distribution with specified location (or mean) and scale (decay). The Laplace distribution is similar to the Gaussian/normal distribution, but is sharper at the peak and has fatter tails. It represents the difference between two independent, identically distributed exponential random variables. .. note:: New code should use the ``laplace`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- loc : float or array_like of floats, optional The position, :math:`\mu`, of the distribution peak. Default is 0. scale : float or array_like of floats, optional :math:`\lambda`, the exponential decay. Default is 1. Must be non- negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``loc`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(loc, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Laplace distribution. See Also -------- Generator.laplace: which should be used for new code. Notes ----- It has the probability density function .. math:: f(x; \mu, \lambda) = \frac{1}{2\lambda} \exp\left(-\frac{|x - \mu|}{\lambda}\right). The first law of Laplace, from 1774, states that the frequency of an error can be expressed as an exponential function of the absolute magnitude of the error, which leads to the Laplace distribution. For many problems in economics and health sciences, this distribution seems to model the data better than the standard Gaussian distribution. References ---------- .. [1] Abramowitz, M. and Stegun, I. A. (Eds.). "Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing," New York: Dover, 1972. .. [2] Kotz, Samuel, et. al. "The Laplace Distribution and Generalizations, " Birkhauser, 2001. .. [3] Weisstein, Eric W. "Laplace Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/LaplaceDistribution.html .. [4] Wikipedia, "Laplace distribution", https://en.wikipedia.org/wiki/Laplace_distribution Examples -------- Draw samples from the distribution >>> loc, scale = 0., 1. >>> s = np.random.laplace(loc, scale, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, density=True) >>> x = np.arange(-8., 8., .01) >>> pdf = np.exp(-abs(x-loc)/scale)/(2.*scale) >>> plt.plot(x, pdf) Plot Gaussian for comparison: >>> g = (1/(scale * np.sqrt(2 * np.pi)) * ... np.exp(-(x - loc)**2 / (2 * scale**2))) >>> plt.plot(x,g) laplacelamlkwargskeykappaitemsizeitemissubdtypeisscalarisnativeisnanisfiniteintpint8int64int32int16index__import__ignoreid hypergeometric(ngood, nbad, nsample, size=None) Draw samples from a Hypergeometric distribution. Samples are drawn from a hypergeometric distribution with specified parameters, `ngood` (ways to make a good selection), `nbad` (ways to make a bad selection), and `nsample` (number of items sampled, which is less than or equal to the sum ``ngood + nbad``). .. note:: New code should use the ``hypergeometric`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- ngood : int or array_like of ints Number of ways to make a good selection. Must be nonnegative. nbad : int or array_like of ints Number of ways to make a bad selection. Must be nonnegative. nsample : int or array_like of ints Number of items sampled. Must be at least 1 and at most ``ngood + nbad``. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if `ngood`, `nbad`, and `nsample` are all scalars. Otherwise, ``np.broadcast(ngood, nbad, nsample).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized hypergeometric distribution. Each sample is the number of good items within a randomly selected subset of size `nsample` taken from a set of `ngood` good items and `nbad` bad items. See Also -------- scipy.stats.hypergeom : probability density function, distribution or cumulative density function, etc. Generator.hypergeometric: which should be used for new code. Notes ----- The probability density for the Hypergeometric distribution is .. math:: P(x) = \frac{\binom{g}{x}\binom{b}{n-x}}{\binom{g+b}{n}}, where :math:`0 \le x \le n` and :math:`n-b \le x \le g` for P(x) the probability of ``x`` good results in the drawn sample, g = `ngood`, b = `nbad`, and n = `nsample`. Consider an urn with black and white marbles in it, `ngood` of them are black and `nbad` are white. If you draw `nsample` balls without replacement, then the hypergeometric distribution describes the distribution of black balls in the drawn sample. Note that this distribution is very similar to the binomial distribution, except that in this case, samples are drawn without replacement, whereas in the Binomial case samples are drawn with replacement (or the sample space is infinite). As the sample space becomes large, this distribution approaches the binomial. References ---------- .. [1] Lentner, Marvin, "Elementary Applied Statistics", Bogden and Quigley, 1972. .. [2] Weisstein, Eric W. "Hypergeometric Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/HypergeometricDistribution.html .. [3] Wikipedia, "Hypergeometric distribution", https://en.wikipedia.org/wiki/Hypergeometric_distribution Examples -------- Draw samples from the distribution: >>> ngood, nbad, nsamp = 100, 2, 10 # number of good, number of bad, and number of samples >>> s = np.random.hypergeometric(ngood, nbad, nsamp, 1000) >>> from matplotlib.pyplot import hist >>> hist(s) # note that it is very unlikely to grab both bad items Suppose you have an urn with 15 white and 15 black marbles. If you pull 15 marbles at random, how likely is it that 12 or more of them are one color? >>> s = np.random.hypergeometric(15, 15, 15, 100000) >>> sum(s>=12)/100000. + sum(s<=3)/100000. # answer = 0.003 ... pretty unlikely! hypergeometrichighhas_gauss gumbel(loc=0.0, scale=1.0, size=None) Draw samples from a Gumbel distribution. Draw samples from a Gumbel distribution with specified location and scale. For more information on the Gumbel distribution, see Notes and References below. .. note:: New code should use the ``gumbel`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- loc : float or array_like of floats, optional The location of the mode of the distribution. Default is 0. scale : float or array_like of floats, optional The scale parameter of the distribution. Default is 1. Must be non- negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``loc`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(loc, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Gumbel distribution. See Also -------- scipy.stats.gumbel_l scipy.stats.gumbel_r scipy.stats.genextreme weibull Generator.gumbel: which should be used for new code. Notes ----- The Gumbel (or Smallest Extreme Value (SEV) or the Smallest Extreme Value Type I) distribution is one of a class of Generalized Extreme Value (GEV) distributions used in modeling extreme value problems. The Gumbel is a special case of the Extreme Value Type I distribution for maximums from distributions with "exponential-like" tails. The probability density for the Gumbel distribution is .. math:: p(x) = \frac{e^{-(x - \mu)/ \beta}}{\beta} e^{ -e^{-(x - \mu)/ \beta}}, where :math:`\mu` is the mode, a location parameter, and :math:`\beta` is the scale parameter. The Gumbel (named for German mathematician Emil Julius Gumbel) was used very early in the hydrology literature, for modeling the occurrence of flood events. It is also used for modeling maximum wind speed and rainfall rates. It is a "fat-tailed" distribution - the probability of an event in the tail of the distribution is larger than if one used a Gaussian, hence the surprisingly frequent occurrence of 100-year floods. Floods were initially modeled as a Gaussian process, which underestimated the frequency of extreme events. It is one of a class of extreme value distributions, the Generalized Extreme Value (GEV) distributions, which also includes the Weibull and Frechet. The function has a mean of :math:`\mu + 0.57721\beta` and a variance of :math:`\frac{\pi^2}{6}\beta^2`. References ---------- .. [1] Gumbel, E. J., "Statistics of Extremes," New York: Columbia University Press, 1958. .. [2] Reiss, R.-D. and Thomas, M., "Statistical Analysis of Extreme Values from Insurance, Finance, Hydrology and Other Fields," Basel: Birkhauser Verlag, 2001. Examples -------- Draw samples from the distribution: >>> mu, beta = 0, 0.1 # location and scale >>> s = np.random.gumbel(mu, beta, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, density=True) >>> plt.plot(bins, (1/beta)*np.exp(-(bins - mu)/beta) ... * np.exp( -np.exp( -(bins - mu) /beta) ), ... linewidth=2, color='r') >>> plt.show() Show how an extreme value distribution can arise from a Gaussian process and compare to a Gaussian: >>> means = [] >>> maxima = [] >>> for i in range(0,1000) : ... a = np.random.normal(mu, beta, 1000) ... means.append(a.mean()) ... maxima.append(a.max()) >>> count, bins, ignored = plt.hist(maxima, 30, density=True) >>> beta = np.std(maxima) * np.sqrt(6) / np.pi >>> mu = np.mean(maxima) - 0.57721*beta >>> plt.plot(bins, (1/beta)*np.exp(-(bins - mu)/beta) ... * np.exp(-np.exp(-(bins - mu)/beta)), ... linewidth=2, color='r') >>> plt.plot(bins, 1/(beta * np.sqrt(2 * np.pi)) ... * np.exp(-(bins - mu)**2 / (2 * beta**2)), ... linewidth=2, color='g') >>> plt.show() gumbelgreaterget_state and legacy can only be used with the MT19937 BitGenerator. To silence this warning, set `legacy` to False.get_stateget geometric(p, size=None) Draw samples from the geometric distribution. Bernoulli trials are experiments with one of two outcomes: success or failure (an example of such an experiment is flipping a coin). The geometric distribution models the number of trials that must be run in order to achieve success. It is therefore supported on the positive integers, ``k = 1, 2, ...``. The probability mass function of the geometric distribution is .. math:: f(k) = (1 - p)^{k - 1} p where `p` is the probability of success of an individual trial. .. note:: New code should use the ``geometric`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- p : float or array_like of floats The probability of success of an individual trial. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``p`` is a scalar. Otherwise, ``np.array(p).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized geometric distribution. See Also -------- Generator.geometric: which should be used for new code. Examples -------- Draw ten thousand values from the geometric distribution, with the probability of an individual success equal to 0.35: >>> z = np.random.geometric(p=0.35, size=10000) How many trials succeeded after a single run? >>> (z == 1).sum() / 10000. 0.34889999999999999 #random geometricgauss gamma(shape, scale=1.0, size=None) Draw samples from a Gamma distribution. Samples are drawn from a Gamma distribution with specified parameters, `shape` (sometimes designated "k") and `scale` (sometimes designated "theta"), where both parameters are > 0. .. note:: New code should use the ``gamma`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- shape : float or array_like of floats The shape of the gamma distribution. Must be non-negative. scale : float or array_like of floats, optional The scale of the gamma distribution. Must be non-negative. Default is equal to 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``shape`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(shape, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized gamma distribution. See Also -------- scipy.stats.gamma : probability density function, distribution or cumulative density function, etc. Generator.gamma: which should be used for new code. Notes ----- The probability density for the Gamma distribution is .. math:: p(x) = x^{k-1}\frac{e^{-x/\theta}}{\theta^k\Gamma(k)}, where :math:`k` is the shape and :math:`\theta` the scale, and :math:`\Gamma` is the Gamma function. The Gamma distribution is often used to model the times to failure of electronic components, and arises naturally in processes for which the waiting times between Poisson distributed events are relevant. References ---------- .. [1] Weisstein, Eric W. "Gamma Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/GammaDistribution.html .. [2] Wikipedia, "Gamma distribution", https://en.wikipedia.org/wiki/Gamma_distribution Examples -------- Draw samples from the distribution: >>> shape, scale = 2., 2. # mean=4, std=2*sqrt(2) >>> s = np.random.gamma(shape, scale, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> import scipy.special as sps # doctest: +SKIP >>> count, bins, ignored = plt.hist(s, 50, density=True) >>> y = bins**(shape-1)*(np.exp(-bins/scale) / # doctest: +SKIP ... (sps.gamma(shape)*scale**shape)) >>> plt.plot(bins, y, linewidth=2, color='r') # doctest: +SKIP >>> plt.show() gammaformatfloat64finfo f(dfnum, dfden, size=None) Draw samples from an F distribution. Samples are drawn from an F distribution with specified parameters, `dfnum` (degrees of freedom in numerator) and `dfden` (degrees of freedom in denominator), where both parameters must be greater than zero. The random variate of the F distribution (also known as the Fisher distribution) is a continuous probability distribution that arises in ANOVA tests, and is the ratio of two chi-square variates. .. note:: New code should use the ``f`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- dfnum : float or array_like of floats Degrees of freedom in numerator, must be > 0. dfden : float or array_like of float Degrees of freedom in denominator, must be > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``dfnum`` and ``dfden`` are both scalars. Otherwise, ``np.broadcast(dfnum, dfden).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Fisher distribution. See Also -------- scipy.stats.f : probability density function, distribution or cumulative density function, etc. Generator.f: which should be used for new code. Notes ----- The F statistic is used to compare in-group variances to between-group variances. Calculating the distribution depends on the sampling, and so it is a function of the respective degrees of freedom in the problem. The variable `dfnum` is the number of samples minus one, the between-groups degrees of freedom, while `dfden` is the within-groups degrees of freedom, the sum of the number of samples in each group minus the number of groups. References ---------- .. [1] Glantz, Stanton A. "Primer of Biostatistics.", McGraw-Hill, Fifth Edition, 2002. .. [2] Wikipedia, "F-distribution", https://en.wikipedia.org/wiki/F-distribution Examples -------- An example from Glantz[1], pp 47-40: Two groups, children of diabetics (25 people) and children from people without diabetes (25 controls). Fasting blood glucose was measured, case group had a mean value of 86.1, controls had a mean value of 82.2. Standard deviations were 2.09 and 2.49 respectively. Are these data consistent with the null hypothesis that the parents diabetic status does not affect their children's blood glucose levels? Calculating the F statistic from the data gives a value of 36.01. Draw samples from the distribution: >>> dfnum = 1. # between group degrees of freedom >>> dfden = 48. # within groups degrees of freedom >>> s = np.random.f(dfnum, dfden, 1000) The lower bound for the top 1% of the samples is : >>> np.sort(s)[-10] 7.61988120985 # random So there is about a 1% chance that the F statistic will exceed 7.62, the measured value is 36, so the null hypothesis is rejected at the 1% level. fexponential__exit__equaleps__enter__empty_likeemptydtypedoubledot dirichlet(alpha, size=None) Draw samples from the Dirichlet distribution. Draw `size` samples of dimension k from a Dirichlet distribution. A Dirichlet-distributed random variable can be seen as a multivariate generalization of a Beta distribution. The Dirichlet distribution is a conjugate prior of a multinomial distribution in Bayesian inference. .. note:: New code should use the ``dirichlet`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- alpha : sequence of floats, length k Parameter of the distribution (length ``k`` for sample of length ``k``). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n)``, then ``m * n * k`` samples are drawn. Default is None, in which case a vector of length ``k`` is returned. Returns ------- samples : ndarray, The drawn samples, of shape ``(size, k)``. Raises ------- ValueError If any value in ``alpha`` is less than or equal to zero See Also -------- Generator.dirichlet: which should be used for new code. Notes ----- The Dirichlet distribution is a distribution over vectors :math:`x` that fulfil the conditions :math:`x_i>0` and :math:`\sum_{i=1}^k x_i = 1`. The probability density function :math:`p` of a Dirichlet-distributed random vector :math:`X` is proportional to .. math:: p(x) \propto \prod_{i=1}^{k}{x^{\alpha_i-1}_i}, where :math:`\alpha` is a vector containing the positive concentration parameters. The method uses the following property for computation: let :math:`Y` be a random vector which has components that follow a standard gamma distribution, then :math:`X = \frac{1}{\sum_{i=1}^k{Y_i}} Y` is Dirichlet-distributed References ---------- .. [1] David McKay, "Information Theory, Inference and Learning Algorithms," chapter 23, http://www.inference.org.uk/mackay/itila/ .. [2] Wikipedia, "Dirichlet distribution", https://en.wikipedia.org/wiki/Dirichlet_distribution Examples -------- Taking an example cited in Wikipedia, this distribution can be used if one wanted to cut strings (each of initial length 1.0) into K pieces with different lengths, where each piece had, on average, a designated average length, but allowing some variation in the relative sizes of the pieces. >>> s = np.random.dirichlet((10, 5, 3), 20).transpose() >>> import matplotlib.pyplot as plt >>> plt.barh(range(20), s[0]) >>> plt.barh(range(20), s[1], left=s[0], color='g') >>> plt.barh(range(20), s[2], left=s[0]+s[1], color='r') >>> plt.title("Lengths of Strings") dirichletdfnumdfdendfcumsumcovariance is not positive-semidefinite.cov must be 2 dimensional and squarecovcount_nonzerocopycompatcollections.abccline_in_traceback__class__ choice(a, size=None, replace=True, p=None) Generates a random sample from a given 1-D array .. versionadded:: 1.7.0 .. note:: New code should use the ``choice`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- a : 1-D array-like or int If an ndarray, a random sample is generated from its elements. If an int, the random sample is generated as if it were ``np.arange(a)`` size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. replace : boolean, optional Whether the sample is with or without replacement. Default is True, meaning that a value of ``a`` can be selected multiple times. p : 1-D array-like, optional The probabilities associated with each entry in a. If not given, the sample assumes a uniform distribution over all entries in ``a``. Returns ------- samples : single item or ndarray The generated random samples Raises ------ ValueError If a is an int and less than zero, if a or p are not 1-dimensional, if a is an array-like of size 0, if p is not a vector of probabilities, if a and p have different lengths, or if replace=False and the sample size is greater than the population size See Also -------- randint, shuffle, permutation Generator.choice: which should be used in new code Notes ----- Setting user-specified probabilities through ``p`` uses a more general but less efficient sampler than the default. The general sampler produces a different sample than the optimized sampler even if each element of ``p`` is 1 / len(a). Sampling random rows from a 2-D array is not possible with this function, but is possible with `Generator.choice` through its ``axis`` keyword. Examples -------- Generate a uniform random sample from np.arange(5) of size 3: >>> np.random.choice(5, 3) array([0, 3, 4]) # random >>> #This is equivalent to np.random.randint(0,5,3) Generate a non-uniform random sample from np.arange(5) of size 3: >>> np.random.choice(5, 3, p=[0.1, 0, 0.3, 0.6, 0]) array([3, 3, 0]) # random Generate a uniform random sample from np.arange(5) of size 3 without replacement: >>> np.random.choice(5, 3, replace=False) array([3,1,0]) # random >>> #This is equivalent to np.random.permutation(np.arange(5))[:3] Generate a non-uniform random sample from np.arange(5) of size 3 without replacement: >>> np.random.choice(5, 3, replace=False, p=[0.1, 0, 0.3, 0.6, 0]) array([2, 3, 0]) # random Any of the above can be repeated with an arbitrary array-like instead of just integers. For instance: >>> aa_milne_arr = ['pooh', 'rabbit', 'piglet', 'Christopher'] >>> np.random.choice(aa_milne_arr, 5, p=[0.5, 0.1, 0.1, 0.3]) array(['pooh', 'pooh', 'pooh', 'Christopher', 'piglet'], # random dtype=' 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``df`` is a scalar. Otherwise, ``np.array(df).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized chi-square distribution. Raises ------ ValueError When `df` <= 0 or when an inappropriate `size` (e.g. ``size=-1``) is given. See Also -------- Generator.chisquare: which should be used for new code. Notes ----- The variable obtained by summing the squares of `df` independent, standard normally distributed random variables: .. math:: Q = \sum_{i=0}^{\mathtt{df}} X^2_i is chi-square distributed, denoted .. math:: Q \sim \chi^2_k. The probability density function of the chi-squared distribution is .. math:: p(x) = \frac{(1/2)^{k/2}}{\Gamma(k/2)} x^{k/2 - 1} e^{-x/2}, where :math:`\Gamma` is the gamma function, .. math:: \Gamma(x) = \int_0^{-\infty} t^{x - 1} e^{-t} dt. References ---------- .. [1] NIST "Engineering Statistics Handbook" https://www.itl.nist.gov/div898/handbook/eda/section3/eda3666.htm Examples -------- >>> np.random.chisquare(2,4) array([ 1.89920014, 9.00867716, 3.13710533, 5.62318272]) # random chisquarecheck_valid must equal 'warn', 'raise', or 'ignore'check_validcastingcapsulecan only re-seed a MT19937 BitGenerator bytes(length) Return random bytes. .. note:: New code should use the ``bytes`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- length : int Number of random bytes. Returns ------- out : bytes String of length `length`. See Also -------- Generator.bytes: which should be used for new code. Examples -------- >>> np.random.bytes(10) b' eh\x85\x022SZ\xbf\xa4' #random bytesbool_bit_generator binomial(n, p, size=None) Draw samples from a binomial distribution. Samples are drawn from a binomial distribution with specified parameters, n trials and p probability of success where n an integer >= 0 and p is in the interval [0,1]. (n may be input as a float, but it is truncated to an integer in use) .. note:: New code should use the ``binomial`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- n : int or array_like of ints Parameter of the distribution, >= 0. Floats are also accepted, but they will be truncated to integers. p : float or array_like of floats Parameter of the distribution, >= 0 and <=1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``n`` and ``p`` are both scalars. Otherwise, ``np.broadcast(n, p).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized binomial distribution, where each sample is equal to the number of successes over the n trials. See Also -------- scipy.stats.binom : probability density function, distribution or cumulative density function, etc. Generator.binomial: which should be used for new code. Notes ----- The probability density for the binomial distribution is .. math:: P(N) = \binom{n}{N}p^N(1-p)^{n-N}, where :math:`n` is the number of trials, :math:`p` is the probability of success, and :math:`N` is the number of successes. When estimating the standard error of a proportion in a population by using a random sample, the normal distribution works well unless the product p*n <=5, where p = population proportion estimate, and n = number of samples, in which case the binomial distribution is used instead. For example, a sample of 15 people shows 4 who are left handed, and 11 who are right handed. Then p = 4/15 = 27%. 0.27*15 = 4, so the binomial distribution should be used in this case. References ---------- .. [1] Dalgaard, Peter, "Introductory Statistics with R", Springer-Verlag, 2002. .. [2] Glantz, Stanton A. "Primer of Biostatistics.", McGraw-Hill, Fifth Edition, 2002. .. [3] Lentner, Marvin, "Elementary Applied Statistics", Bogden and Quigley, 1972. .. [4] Weisstein, Eric W. "Binomial Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/BinomialDistribution.html .. [5] Wikipedia, "Binomial distribution", https://en.wikipedia.org/wiki/Binomial_distribution Examples -------- Draw samples from the distribution: >>> n, p = 10, .5 # number of trials, probability of each trial >>> s = np.random.binomial(n, p, 1000) # result of flipping a coin 10 times, tested 1000 times. A real world example. A company drills 9 wild-cat oil exploration wells, each with an estimated probability of success of 0.1. All nine wells fail. What is the probability of that happening? Let's do 20,000 trials of the model, and count the number that generate zero positive results. >>> sum(np.random.binomial(9, 0.1, 20000) == 0)/20000. # answer = 0.38885, or 38%. binomialbetabatol at 0x{:X}astypeasarrayarrayargsarangeanyalpha <= 0alphaallclose__all__alladda must be greater than 0 unless no samples are takena must be 1-dimensional or an integera must be 1-dimensional'a' cannot be empty unless no samples are taken'a' and 'p' must have same sizea)(ValueErrorUserWarningUnsupported dtype %r for randintTypeErrorThis function is deprecated. Please call randint({low}, {high} + 1) insteadThis function is deprecated. Please call randint(1, {low} + 1) insteadTShuffling a one dimensional array subclass containing objects gives incorrect results for most array subclasses. Please us the new random number API instead: https://numpy.org/doc/stable/reference/random/index.html The new API fixes this issue. This version will not be fixed due to stability guarantees of the API.SequenceRuntimeWarningRange exceeds valid boundsRandomState.zipf (line 3599)RandomState.weibull (line 2393)RandomState.wald (line 3096)RandomState.vonmises (line 2203)RandomState.uniform (line 1014)RandomState.triangular (line 3172)RandomState.tomaxint (line 588)RandomState.standard_t (line 2089)RandomState.standard_normal (line 1341)RandomState.standard_gamma (line 1513)RandomState.standard_exponential (line 546)RandomState.standard_cauchy (line 2016)RandomState.shuffle (line 4422)RandomState.seed (line 224)RandomState.rayleigh (line 3020)RandomState.random_sample (line 374)RandomState.random_integers (line 1245)RandomState.randn (line 1181)RandomState.randint (line 646)RandomState.rand (line 1137)RandomState.power (line 2496)RandomState.poisson (line 3517)RandomState.permutation (line 4543)RandomState.pareto (line 2291)RandomState.normal (line 1406)RandomState.noncentral_f (line 1769)RandomState.noncentral_chisquare (line 1929)RandomState.negative_binomial (line 3431)RandomState.multivariate_normal (line 3967)RandomState.multinomial (line 4141)RandomState.logseries (line 3879)RandomState.lognormal (line 2905)RandomState.logistic (line 2820)RandomState.laplace (line 2604)RandomState.hypergeometric (line 3746)RandomState.gumbel (line 2697)RandomState.geometric (line 3685)RandomState.gamma (line 1593)RandomState.f (line 1676)RandomState.dirichlet (line 4274)RandomState.choice (line 807)RandomState.chisquare (line 1854)RandomState.bytes (line 771)RandomState.binomial (line 3280)RandomStateProviding a dtype with a non-native byteorder is not supported. If you require platform-independent byteorder, call byteswap when required. In future version, providing byteorder will raise a ValueErrorOverflowErrorNegative dimensions are not allowedMT19937_MT19937Invalid bit generator. The bit generator must be instantized.IndexErrorImportErrorFewer non-zero entries in p than sizeDeprecationWarningCannot take a larger sample than population when 'replace=False':0yE>qh??name '%U' is not definednumpy/random/mtrand.c%s (%s:%d)__pyx_capi__need more than %zd value%.1s to unpack%.200s() keywords must be strings%s() got an unexpected keyword argument '%U'%.200s.%.200s is not a type object%.200s.%.200s size changed, may indicate binary incompatibility. Expected %zd from C header, got %zd from PyObject%s.%s size changed, may indicate binary incompatibility. Expected %zd from C header, got %zd from PyObject%.200s does not export expected C variable %.200sC variable %.200s.%.200s has wrong signature (expected %.500s, got %.500s)%.200s does not export expected C function %.200sC function %.200s.%.200s has wrong signature (expected %.500s, got %.500s)__int__ returned non-int (type %.200s). The ability to return an instance of a strict subclass of int is deprecated, and may be removed in a future version of Python.__%.4s__ returned non-%.4s (type %.200s)value too large to convert to intcalling %R should have returned an instance of BaseException, not %Rraise: exception class must be a subclass of BaseException'%.200s' object is not subscriptablecannot fit '%.200s' into an index-sized integer%s() got multiple values for keyword argument '%U'O@ONNN\PtP?x,>N>>q>>>j>>k>>Π>>F>>>'>\>#>u>J>*>_F>d>+>$>w>>>JK>y>|>iݿ>>I>;>ʾ>t>5<> ~>>>O>>>~3>T>ե>(>g~>ՠ>G/>>>F>J> >:n>bԓ>Q<>>x>~>>>^>Ј>D>l>1>>%>\D>@|>?y>Bv>Hs>Qp>#^m>mj>|g>md>a>^>$[> Y>=3V>[S>P>M>J>~H>UE>B>?>=>S:>7>"4>=22>T/>d,>m+*>m'>c$>N?">,>>m>t>F>>1*> > >Y>>ʗ>>I=_={==^==&=_=g=='0===P6=˙=\= s==d= =yo=/=6=.=fЍ=x=i'=܀=a1y=p=xIh=_==W=TO=G=>=N6=.=&===-H==<א<̀<<<.~2~~r~դ~Ƭ~N~u~C~~~~k~~~~~~t~~~6 < :#%](*.-z/13579;=?EABD:FGNIJ8LMNLPQR T=UdVWXYZ[\]^~__`;abbcod.eefLggh~~7~~/~7~~ ~ ~w~G]~>~Y~,}6}b}|O|06{xI4_h2z3+3@3aQ3i`3{am3Ay3i3*353=3r333|ϡ3ڍ3+333^33׶3iż3-¿3c3%3uY3<3L3gv3;3k3-3$3!333P3P33<3p~3չ3^3J3I34V?4=3@4A4A4qB4C4D4udE4-CF4K"G4H4H41I4J4vK4\fL4HM4+N4aO4O4bP4ٽQ4R4ԊS4crT4ZU4CV4-W4ZX4Y4UY4Z4[4(\4_]4^4_4C`4va4alb40cc47[d4~Te4Of4Jg42Hh4Fi4Fj4Hk4Kl4MPm4Vn4^o48hp4sq4r4s4 t4u4v4Cw4x4 z42{40S|4u}4~44v4@ 4L4>4ق4v444lV44R4F44p4 I44"4_44Ќ4l4L4`4ԏ4坐4y4ݖ4%44r&4k44(4444.4Q4N4t44\۶4H94̻4p44~X4w4p_4~444?yjD?l[T?w'??o?Wp?xI?-3?x^j??N?R:e?4:>?l?*?%z?PՋt?4?e;?$"?zaWF}?Gz‘B?Oq1? OU?ߺH?7a?nV,? K?Xhw?հ<?Vp\?m?)?zP?ZcX?*;Q^?#*'g? U7?e&$ ?jJo?\Ȭ)?L&?FS?leZ&?g ?NIO??xRr!?P_hy?y6IJO?_5%?[X~?1>?bU?+À?PX?5:pɗ0?8d?;U?J?͓?)m?ېZ]G?/|!? ?iT??Wq?PF9 ?ߓ^??ۮY?3???i?Z8o? O5?ٸ?P?R9?igP?La;?L?!ވ?%o?{7=8?Ҁt?DvC?6?=p\?;So&?mj?W?j?$O?z5Ѽ?Ҏ?C|P?yh|?%H?/ZM?f!w;??>ǭ?MAz?G?y?.?P9կ?TT}?g4K?#$O? Y?BM?6C;?B"_U?~t$?œ߉?52?Ҙl'?DɤT?<(i?qE8 ? Uī?OQM?o^?Sq͒?Gط5?zx?1zd}?:R!?Wg?~& ~k?=~-2?ZҿҶ?'|j_]?it?[?8R?uqb?#h?z|J?G~`?\!>?GF?vJ?l󈬚?5hȩmE?㭍?-l ?uG?1i%?調?M?e*|?zè?^V?4<%F?B}u?c-@c?n? R=?Kr?*}T#?,"k>?R) ?K{o?vaӽ?命8? t;I_? h?3xk?3Ӻ?b3?vZ9S?LJisk?M$a.?ftW?+ ?"@|?&#?p>_?1fҲ? DE?} ?/?%,?0?5nl+,&?QG?b. ?,*(>?p_8?cU)?h*?'wާ?dИۦ?ԭ<ڥ?]']ۤ?ݣ?=|?j?.?ĥׁ?u? ̓0?"NR? y? ڥ?d֔?^8 ?0`4I?IrO*?O'?x A?B?/)?7h`|?] ٨v?p?gC_e?T?j%= BT ~Q~U~KD Ga7\%aFOaSuzpD(|Wc %WM$ t`K[oT`gtSwf# Wl`0H7[z1z(zK^2#9MM0MFPrOxS왎2ȩn{TH,ҭ^p .]M[\}r;/4d6dcNQp.t@e$oX%L(xo6xu{fUY>9>{ppCBwS(:5^dܓAN}8) YfHqն&|s f2,2Ztզޗ .n ZR'ӯB)[l@u Pҍ'TȈt(5wI'L/$;nXMØT`OArW,+jtȳRfARnqӊ<KZW$eKs) 4<=>)G'QA@Y.(5bX jz>lq{2Xx{~JH҄Cc`Qz%~ )Q\HsrUb'Bkq-hnק Ψ;3Kd)P^٨Tv$Hx"$ 5..&$ŗ: L< Ŀk<4xV<=A[<'?}y<NG<~;[ ?A?Į?"?ʝ?G??i>l>7>>*J>>>^>>F>>7P>>K{>>>u>;->>>|>eO>4(>8>L>N>ȿ>>>>·>ε>߳>>>0>U>~>|>ަ>Y>IP>w>ҟ>B>e>)>~>_>>C>{>>J>﷍>X(>'>N>Í>x >b>x>>!>^}>;z>Хw>@t>wr>byl>i>g>Sd>3a>^>]\>&Y>z)W>T>P R>O>L>5~J>3H>E>nC>@>VK>>;>9>07>4>2>>0>p->+>s)>7'>%>"> >s>L>*>= >T>T>4>>y >ϣ > >>L>>l>=+==0 =C==8==hp==0== =n==|===;=ں=Z=o,=ް=ߗ=.W===%==r=W= C=4=M,=4*=D.=y8=H=~=x=3E3lM3FT3/[3b3i34p3fw3&~3[3B3ψ3g37!3>3T3d3n3r3Fq3j3_31P3r<3$3k 33ȸ3q3|{3P3#3C333dY3"3+3®3r35333x37333p3-33731b33l44(44h4C44 4` 4M 47 44?4nB44L4 i4a4T044542`44p. 4!4"4i$4%4@'4t(4>*4+4,4j.4/4'P1424):4454&)7484c:4;4$=4+>4@4A4KC4vD4B(F4G4:I4J4rTL4M4GuO4Q4R44T4U4EiW4Y4 Z4G\4]4_4:a4b4d4Bf4\g4ji4bk4m4n4p4\r4}"t4Yu4Hw4[y4X{46.}4 4q4a4]S4F4N<434,4+(4{%4$4o&4,*4'04m84 C4P4_4q474{4w4>ԕ44s4<4d444$4 (4a44lߢ4$4l44x 4_444{4 4EP4±4{:4귴4);4nķ4S444<*4տ44A4.44ע44f4RW4R4*Y4Fk44δ444444g44k4<444y44u4_45555@5ó 5 5]5^555q5v 5!5%5V*5s/5;S55:<5D5NO5^5Nv5?7E?P?'{{?*!?bv?mU?9U1T?/v?x]?&1$-?~ n?cK[!?I?\Omg?f?uLi=?sڂl?x?Qf?ij?%ᨯC?+?Dܻ?z?cE#;?^E#?$O?2m?P"K?>?{s?%;?omo?3;?J9?++?*T[?};1s?HeC?$`?vE!=?ſ-r?MBц?K=?Q}6Ei?7u? !?z}k? ~?@?`x?*?8? Qi?oTC?_(4?ָ?@je?!u v?7Zi?{ ?I?]T?9]??}?8aD:?Yζi?Ɲҷ?r^sSw?ꍰ07?d>[?%۹? Ə{?'HB>?vX#?l1&?:l?磽!O?ލ?&?ڋ?ﾀ+ j?A1?N0Z?0H?}G?(V?5$1!?pB9 ?b"FS?)vEW(?vG}rO?~ /? {^?Z? ?ބS?i"?lR?3Sn?>N?Ґ]b?,|y2?jG>?TLҫ?~>\O??@YH?/֎@?9O"H?>?1 7?8?Ox?]4?5D9g?r|?>ܸ$8? [B/?I䠟?O?y%d?bPޱ?c?PR?j?F}?9(Q1?c?(ڦ^w?0U^Q?1j?T ξ?x.BTv?Imb.??6YJ?)ِ?\C}?%d?w?SN?эv?pa?,Q&?@oű?SuFe?PV?;?I?viׯ?4D?.g?X1Iα?Jy?!dJ?پz?j»?8G;?L|{ʎ?mwn?k9:9??Ry?A&E?U?Ŗ<?k&_?G??~#? V#?_?S?Q| z? Y&?$?htQz? 3Tݜ?pXP?N梚?H*g?gS(u?1c?w@rT?Q=I?QA?]1%? RD?lj?W'n?-BU؊?h?t4? n?boQ?qvi?_)N?]tQW}?6H#z? 67w?"Ηs?C@Wi=q?ḰXl?f?$ka?%> T+Y? O?K 2=?Ɨ$'R~1[} NVeΙVn6nvK zicp%E tQ)2U1WQ9Lin?23F:L"3\LQ V f[_rWDdx h+*k2=Ko:qr Mu\x?A{FS~8;b=ZV`bBtu9=JE>XدGwdO 8cx AFẙi&zqVYםΡag6 X83:뇡koɣj_ۤ| Mg^ݧt|Ψ_ΓXp2X^ttH蟿W;ޭl~$\z[߁İPp:J+N!X ɦ֬ ᆴX7(. Ɉ?5}h.G{tr&oya=cA/˺DH0⤮<)9O@ᣩTrVj֋@?˷dsI^i@(0߾ta&⊂l1EA1T[n&mi#d)B}QJwt}B < EOvpc/F<Ң"Ae އ0~ Rfq(*QtH3D@M`P}hwx%ƿ8*JG+[EliPIw+ E>ҙ02yΩ4A (Nt.Ȱ--̕^&܌z#;ޖu~g6X .pmF 3n bH޵LaEZvpR(-x_b˿ӰdyQӶVg<7܆ut7$MH𯋉ld"rqտH)݄ /0 wپ}2}K D5z&R cM,}uc?Ѡp5.bJ3ʸT[vv+\[U@ضBi"7oLeiFγ>SR(D2Z> B0$y1gWr-ެ @樫(afoeW-|&aY +M]d<A]X`<+M[Ij<[5q&<.8eG< h#ឪV <;LC%K<ꆭh-DT!@8?5gG@dg?3?r?q?0@0C@$@= ףp=@n?[ m?h|?5?333333 @r?$~?B>٬ @r鷯?Q?Q?9v?(\@ffffff@.@4@x&??UUUUUU?a@X@`@|@@MA>@C~)@ lѿ3 ; @.A-DT! cܥL@9RFߑ?+eG?9B.?ý.@5?4K?N@Si@>Aޓ=?.eB5<;?r1?; ( !P "h .'X/T'0'r4(0<11@ ` p D `l@XТ`DPD`` l(p DTH"Xp%`(X,/Hp3 78:=@ Uz~ p ` PT!`!<"",#$t$A$A %B4%`BP%B%PC%a &zp& }&0&H((@)P)d)@)p)@) *0$*P<*h*`****P+P8+`\++ +@++P,,,D, \,@t,,,-,-P`--- .0X.@..`/d/x// //0@0 01`102 T2222 2@3`303D3 x303@44Pl4444 5 H5!t50"5"5#5$6$46&6&6`(07*7P*7 - 8.80929p39 4 :044:@4H:P4\:`4p:p4:4:4:4:4:5:5:@5;`5$;p58;5L;5`;5t;5;5;5;5;5;6;6< 6<06(<@6<<P6P<`6d<6x<6<6<6<6<6<6<6=7=7,= 7@=07T=7h=7|=08=p8=8=8=8=8=8>8>80>9D>9X> 9l>09>@9>9>:>0;?;P?;h?;? FBB A(D0D@tHIPYHF@s 0D(A BBBC HHaPBXA`AhBpAxABAQ@TL K FBB B(A0A8GPXH`YXFPv 8D0A(B BBBE HFBB E(A0A8D@r 8D0A(B BBBD D\UFIA j BBB  BBH R BBJ H8tFBB B(A0A8DP 8D0A(B BBBH THpFBB E(A0A8DPXH`YXFPr 8A0A(B BBBD X`I FBB B(A0A8GpxHYxFps 8D0A(B BBBH H8TFBB B(D0A8GP1 8D0A(B BBBJ (4v#FBB B(A0D8D_RDA_QAt 8D0A(B BBBE PRF_OA_MB_RB8GLAXBOAEMBGLAd FBB B(D0A8DPi 8A0A(B BBBH  8C0A(B BBBD \LOFBB B(D0A8G 8D0A(B BBBF  WKFX`FBB A(D0D@tHIPYHF@s 0D(A BBBC HHaPBXA`AhBpAxABAQ@t,FBE A(A0GPuX_`BhApAxEFABATPk 0D(A BBBI sXV`RXFP| dFBB B(A0D8G`zhIpYhF`s 8D0A(B BBBF ZhbpBxAABADBFQ`| FBB B(A0D8G`hPpRhF`s 8D0A(B BBBE Rh[pBxFFBFABFQ` FBB B(A0D8G`eh[pBxFFBFABFQ`o 8D0A(B BBBD ^hWpRhF`hWp_hA` FBB E(A0A8GptxEkxFps 8D0A(B BBBF [xXBFAEFABFQp#xW_xApNxV`xGp|8 FBB B(A0D8G`zhIpYhF`s 8D0A(B BBBF ZhbpBxAABADBFQ` FBB B(A0D8G`eh[pBxFFBFABFQ`o 8D0A(B BBBD ^hWpRhF`hWp_hA`pH 4NFBB A(A0G@tHHPYHF@s 0D(A BBBD OHfPBXA`BhBpAxBBAQ@| FBB B(A0D8G`zhIpYhF`s 8D0A(B BBBF ZhbpBxAABADBFQ`< FBB B(A0D8G`eh[pBxFFBFABFQ`o 8D0A(B BBBD ^hWpRhF`hWp_hA`| FBB B(A0D8G`zhIpYhF`s 8D0A(B BBBF ZhbpBxAABADBFQ`|L FBB B(A0D8G`zhIpYhF`s 8D0A(B BBBF ZhbpBxAABADBFQ`| FBB B(A0D8G`zhIpYhF`s 8D0A(B BBBF ZhbpBxAABADBFQ`tL FBE A(A0GPuX[`BhFpAxEFABFQPk 0D(A BBBF sXV`RXFPt8FBE A(A0GPuX[`BhFpAxEFABFQPk 0D(A BBBF sXV`RXFPt<pFBE A(A0GPuX[`BhFpAxEFABFQPk 0D(A BBBF sXV`RXFPtFBE A(A0GPuX[`BhFpAxEFABFQPk 0D(A BBBF sXV`RXFPt,FBB A(A0GPtXL``XFPs 0D(A BBBI UXb`BhApAxBADBFQPHFBB B(A0D8G`eh[pBxFFBFABFQ`o 8D0A(B BBBD ^hWpRhF`hWp_hA`4x`FBB B(A0A8DHeBAFBABBATk 8D0A(B BBBD FSAFBABBARnVRF6%FBB B(A0D8DeBBBBBBTDkFs 8D0A(B BBBF LFEKKBABBAYW_AV`GL[FBB B(D0 (E BBBH  (E BBBC LX^FBB B(D0 (E BBBH  (E BBBC \daFFBB B(A0D8DORFs 8D0A(B BBBC vFBB D(D0GPtXW`RXFPs 0D(A BBBF WX[`FhFpBxFAEFAQP1XW`_XAPlHyFBD A(GPrXL``XFPs (D ABBD SXi`AhApBxADEFAQPt|!FBB D(D0GPzXI`YXFPs 0D(A BBBG SXb`AhApBxADBFATPt0L!FBB D(D0GPzXI`YXFPs 0D(A BBBG SXb`AhApBxADBFATPt!FBB D(D0GPzXI`YXFPs 0D(A BBBG SXb`AhApBxADBFATP &FBB B(A0D8DeFABFBBTDkFs 8D0A(B BBBF HSEGFBBFATwW_AV`GpIFBB B(A0D8G 8D0A(B BBBG SWRF5(W_FphT)FBB B(A0A8J~WRFy 8D0A(B BBBE W_A4EFDD d GBE AAB (`H{ E 4DEAG O CAH I CAK 4|`QEDG [ CAA M CAG \8FBB B(A0D8G IYFD 8A0A(B BBBE Lh;{FBB E(A0D8D 8D0A(B BBBA @dS FBB B(A0G@ 0D(B BBBJ @dV FBB B(A0G@ 0D(B BBBJ 0YET O |8 xBEH A(A0\(D BBB8HUFBB A(D0;(D BBB QFBB B(A0D8G NEFFAAAAAN{NEFFAAAAANE8A0A(B BBB(6Zl<X FBB B(A0D8GWRFy 8D0A(B BBBC 9W_ADxFFB B(A0A8DP8A0A(B BBBL8y9;FBB B(D0A8D0 8D0A(B BBBF D(p8XIKED D(F0](A ABBG,H c<lBEB D(D0q (D BBBD ܵ3HjH U(0 EG Z AI XA$\pqFFG0MAKȶVEG HA( EG0_ AD  AA (H _3EG ]A 4MO0 EA ,P@ I z E @PEG@ EF x EC 4 AK D AE XH U`HQhWEG }A EG _ AE  4B\ ] l+H b8 CH nP HILh ĽFAD D@  AAFO j  AAFN  AAF $ EG@ DD , EDP AG ^ AI  !%0 !tFEG e EBB lEBHT!|FHA D(G0s (D ABBJ N(D ABB!0\!8FED D(D0| (D ABBD Q (D DBBC D (A ABBE 4"zFAD  ABC D DBF `L"FBE A(A0V (D BBBA j (A EBBD I (D BBBA ("|xFED ^ EBD ("FAG U DBG L#DTFEA A(D@< (A ABBE m (A ABBE X#T58l#FED D(GP (A ABBI #$50#P\FAA D0T  AABD @#|FAA D0`  AABH   FABB @4$FEE A(A0GP 0A(A BBBI Lx$QFHE A(D0JHhn0D(A BBBd$DFBB E(D0D8F`N 8A0A(B BBBA  8A0A(B BBBL D0%URFEL E(A0A8E@8A0A(B BBBDx%RFEL E(A0A8E@8A0A(B BBB %Fz H p H <%hED } AI G AH V AI G AC $&yD F AA (H&XFD P EA I EB t&)K]& 8& YKED D(F0p(A ABBD&0KJ&8KJ '@KJ$'H 08'DzAD@ EAH `Hl'FBB E(A0D8G[ 8D0A(B BBBE l'dFBB A(A0Gpp 0D(A BBBB V 0D(A BBBG / 0D(A BBBA 4(( EO r AI P AG D AE `([EG AA(*MX(0 L(, FFG E(D0A8J 8A0A(E BBBG 8)*FED A(G`S (D ABBE (<)Qa N n J L L A0h)oa H nJHA G L)_H V()ZFAG@CFB) 0)QDD X ABD XH(*dZFBB B(D0A8D@y 8D0A(B BBBE Ht*xFBB B(D0A8DPq 8D0A(B BBBE `*FDB B(A0A8DP  8A0A(B BBBI L 8A0A(D BBBH `$+(FDB B(A0A8DP 8A0A(B BBBG M 8A0A(D BBBG (+dPECD k AAG H+FBB E(D0A8D`h 8A0A(B BBBF , FBB E(D0A8DPh 8A0A(B BBBF  8A0A(B BBBB e 8A0A(B BBBA m8A0A(B BBB`,FDB E(A0A8GPg 8A0A(B BBBE A 8A0A(B BBBE x,IFBE B(A0A8DPR 8K0A(B BBBE _ 8A0A(B BBBG L 8A0A(B BBBJ <t-hyKJE D(E0F(A BBBH\-FIB B(A0D8D`j 8A0A(B BBBA S8A0A(B BBB. (. <. P. d. x. . . .'.'. . $/(/4 ,/0 @/, T/( h/$ |/ / / / / / / 0 0 00 D0 X0l0 0 0 0 0 0 0 0  1  1 41 H1V\1 Vp1l71=11 1 1 1 1 2 $2 82 L2 `2 t2 2LH | A 2vd LH2@FEH H(KP (E ABBG [(A AFB83IKED D(F0^(A ABBFD3(H U\30+H bt3HB\ ]3CH n37H g A 83aKED D(F0r(A ABBJ<3yKED D(G0K(A ABBHH<4PFBE I(H0K8K` 8E0A(B BBBE 84IKED D(F0^(A ABBF@4EG@ EI x EC 1 AN D AE 5H U 5*MX(<50EG0J AQ Ah5HQ5 WEG }A5L(H _5dHI5l3EG ]A 5EG _ AE 86OO j AG l FN DAT P6[EG AA p6,MO0 EA 46:EDP AD V AA f AI 6EG@ DD 6_H Q7D 7P 07L D7H X7D l7@ 7< 7H 7D 7@ 7< 78 74  80  8, 48( H8$ \8 p8 L8X L8 18 58 8 8 8 9 $9 89 L9 `9 t9 9 9 T9 -FBE H(H0K@ 0D(A BBBA m0A(A FBB8: IKED D(F0^(A ABBFHD: FBE B(H0H8KP 8F0A(B BBBD 8:, IKED D(F0^(A ABBF@:@ EG0 DE t DH J AM D AE ;H U(; <; P; d; x; ;; ; ; ; ;9< < ,<HQD< X<Y RU a O Z |<`Y RU a O Z<A< < <HQ< =\ TX b F Z ,=\ TX b F ZP=Bd=(x=4=@HDD I=@8=LHf B H H L N0PD J J L0PD8=Hf B H H L N0PD J J L0PD8>4/$L>P5FDF ^DBt>h7$>8FDG `DB>7$>8FDG `DB>??(?>> np.random.permutation(10) array([1, 7, 4, 3, 0, 9, 2, 5, 8, 6]) # random >>> np.random.permutation([1, 4, 9, 12, 15]) array([15, 1, 9, 4, 12]) # random >>> arr = np.arange(9).reshape((3, 3)) >>> np.random.permutation(arr) array([[6, 7, 8], # random [0, 1, 2], [3, 4, 5]]) shuffle(x) Modify a sequence in-place by shuffling its contents. This function only shuffles the array along the first axis of a multi-dimensional array. The order of sub-arrays is changed but their contents remains the same. .. note:: New code should use the ``shuffle`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- x : ndarray or MutableSequence The array, list or mutable sequence to be shuffled. Returns ------- None See Also -------- Generator.shuffle: which should be used for new code. Examples -------- >>> arr = np.arange(10) >>> np.random.shuffle(arr) >>> arr [1 7 5 2 9 4 3 6 0 8] # random Multi-dimensional arrays are only shuffled along the first axis: >>> arr = np.arange(9).reshape((3, 3)) >>> np.random.shuffle(arr) >>> arr array([[3, 4, 5], # random [6, 7, 8], [0, 1, 2]]) dirichlet(alpha, size=None) Draw samples from the Dirichlet distribution. Draw `size` samples of dimension k from a Dirichlet distribution. A Dirichlet-distributed random variable can be seen as a multivariate generalization of a Beta distribution. The Dirichlet distribution is a conjugate prior of a multinomial distribution in Bayesian inference. .. note:: New code should use the ``dirichlet`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- alpha : sequence of floats, length k Parameter of the distribution (length ``k`` for sample of length ``k``). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n)``, then ``m * n * k`` samples are drawn. Default is None, in which case a vector of length ``k`` is returned. Returns ------- samples : ndarray, The drawn samples, of shape ``(size, k)``. Raises ------- ValueError If any value in ``alpha`` is less than or equal to zero See Also -------- Generator.dirichlet: which should be used for new code. Notes ----- The Dirichlet distribution is a distribution over vectors :math:`x` that fulfil the conditions :math:`x_i>0` and :math:`\sum_{i=1}^k x_i = 1`. The probability density function :math:`p` of a Dirichlet-distributed random vector :math:`X` is proportional to .. math:: p(x) \propto \prod_{i=1}^{k}{x^{\alpha_i-1}_i}, where :math:`\alpha` is a vector containing the positive concentration parameters. The method uses the following property for computation: let :math:`Y` be a random vector which has components that follow a standard gamma distribution, then :math:`X = \frac{1}{\sum_{i=1}^k{Y_i}} Y` is Dirichlet-distributed References ---------- .. [1] David McKay, "Information Theory, Inference and Learning Algorithms," chapter 23, http://www.inference.org.uk/mackay/itila/ .. [2] Wikipedia, "Dirichlet distribution", https://en.wikipedia.org/wiki/Dirichlet_distribution Examples -------- Taking an example cited in Wikipedia, this distribution can be used if one wanted to cut strings (each of initial length 1.0) into K pieces with different lengths, where each piece had, on average, a designated average length, but allowing some variation in the relative sizes of the pieces. >>> s = np.random.dirichlet((10, 5, 3), 20).transpose() >>> import matplotlib.pyplot as plt >>> plt.barh(range(20), s[0]) >>> plt.barh(range(20), s[1], left=s[0], color='g') >>> plt.barh(range(20), s[2], left=s[0]+s[1], color='r') >>> plt.title("Lengths of Strings") multinomial(n, pvals, size=None) Draw samples from a multinomial distribution. The multinomial distribution is a multivariate generalization of the binomial distribution. Take an experiment with one of ``p`` possible outcomes. An example of such an experiment is throwing a dice, where the outcome can be 1 through 6. Each sample drawn from the distribution represents `n` such experiments. Its values, ``X_i = [X_0, X_1, ..., X_p]``, represent the number of times the outcome was ``i``. .. note:: New code should use the ``multinomial`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- n : int Number of experiments. pvals : sequence of floats, length p Probabilities of each of the ``p`` different outcomes. These must sum to 1 (however, the last element is always assumed to account for the remaining probability, as long as ``sum(pvals[:-1]) <= 1)``. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray The drawn samples, of shape *size*, if that was provided. If not, the shape is ``(N,)``. In other words, each entry ``out[i,j,...,:]`` is an N-dimensional value drawn from the distribution. See Also -------- Generator.multinomial: which should be used for new code. Examples -------- Throw a dice 20 times: >>> np.random.multinomial(20, [1/6.]*6, size=1) array([[4, 1, 7, 5, 2, 1]]) # random It landed 4 times on 1, once on 2, etc. Now, throw the dice 20 times, and 20 times again: >>> np.random.multinomial(20, [1/6.]*6, size=2) array([[3, 4, 3, 3, 4, 3], # random [2, 4, 3, 4, 0, 7]]) For the first run, we threw 3 times 1, 4 times 2, etc. For the second, we threw 2 times 1, 4 times 2, etc. A loaded die is more likely to land on number 6: >>> np.random.multinomial(100, [1/7.]*5 + [2/7.]) array([11, 16, 14, 17, 16, 26]) # random The probability inputs should be normalized. As an implementation detail, the value of the last entry is ignored and assumed to take up any leftover probability mass, but this should not be relied on. A biased coin which has twice as much weight on one side as on the other should be sampled like so: >>> np.random.multinomial(100, [1.0 / 3, 2.0 / 3]) # RIGHT array([38, 62]) # random not like: >>> np.random.multinomial(100, [1.0, 2.0]) # WRONG Traceback (most recent call last): ValueError: pvals < 0, pvals > 1 or pvals contains NaNs multivariate_normal(mean, cov, size=None, check_valid='warn', tol=1e-8) Draw random samples from a multivariate normal distribution. The multivariate normal, multinormal or Gaussian distribution is a generalization of the one-dimensional normal distribution to higher dimensions. Such a distribution is specified by its mean and covariance matrix. These parameters are analogous to the mean (average or "center") and variance (standard deviation, or "width," squared) of the one-dimensional normal distribution. .. note:: New code should use the ``multivariate_normal`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- mean : 1-D array_like, of length N Mean of the N-dimensional distribution. cov : 2-D array_like, of shape (N, N) Covariance matrix of the distribution. It must be symmetric and positive-semidefinite for proper sampling. size : int or tuple of ints, optional Given a shape of, for example, ``(m,n,k)``, ``m*n*k`` samples are generated, and packed in an `m`-by-`n`-by-`k` arrangement. Because each sample is `N`-dimensional, the output shape is ``(m,n,k,N)``. If no shape is specified, a single (`N`-D) sample is returned. check_valid : { 'warn', 'raise', 'ignore' }, optional Behavior when the covariance matrix is not positive semidefinite. tol : float, optional Tolerance when checking the singular values in covariance matrix. cov is cast to double before the check. Returns ------- out : ndarray The drawn samples, of shape *size*, if that was provided. If not, the shape is ``(N,)``. In other words, each entry ``out[i,j,...,:]`` is an N-dimensional value drawn from the distribution. See Also -------- Generator.multivariate_normal: which should be used for new code. Notes ----- The mean is a coordinate in N-dimensional space, which represents the location where samples are most likely to be generated. This is analogous to the peak of the bell curve for the one-dimensional or univariate normal distribution. Covariance indicates the level to which two variables vary together. From the multivariate normal distribution, we draw N-dimensional samples, :math:`X = [x_1, x_2, ... x_N]`. The covariance matrix element :math:`C_{ij}` is the covariance of :math:`x_i` and :math:`x_j`. The element :math:`C_{ii}` is the variance of :math:`x_i` (i.e. its "spread"). Instead of specifying the full covariance matrix, popular approximations include: - Spherical covariance (`cov` is a multiple of the identity matrix) - Diagonal covariance (`cov` has non-negative elements, and only on the diagonal) This geometrical property can be seen in two dimensions by plotting generated data-points: >>> mean = [0, 0] >>> cov = [[1, 0], [0, 100]] # diagonal covariance Diagonal covariance means that points are oriented along x or y-axis: >>> import matplotlib.pyplot as plt >>> x, y = np.random.multivariate_normal(mean, cov, 5000).T >>> plt.plot(x, y, 'x') >>> plt.axis('equal') >>> plt.show() Note that the covariance matrix must be positive semidefinite (a.k.a. nonnegative-definite). Otherwise, the behavior of this method is undefined and backwards compatibility is not guaranteed. References ---------- .. [1] Papoulis, A., "Probability, Random Variables, and Stochastic Processes," 3rd ed., New York: McGraw-Hill, 1991. .. [2] Duda, R. O., Hart, P. E., and Stork, D. G., "Pattern Classification," 2nd ed., New York: Wiley, 2001. Examples -------- >>> mean = (1, 2) >>> cov = [[1, 0], [0, 1]] >>> x = np.random.multivariate_normal(mean, cov, (3, 3)) >>> x.shape (3, 3, 2) The following is probably true, given that 0.6 is roughly twice the standard deviation: >>> list((x[0,0,:] - mean) < 0.6) [True, True] # random logseries(p, size=None) Draw samples from a logarithmic series distribution. Samples are drawn from a log series distribution with specified shape parameter, 0 < ``p`` < 1. .. note:: New code should use the ``logseries`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- p : float or array_like of floats Shape parameter for the distribution. Must be in the range (0, 1). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``p`` is a scalar. Otherwise, ``np.array(p).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized logarithmic series distribution. See Also -------- scipy.stats.logser : probability density function, distribution or cumulative density function, etc. Generator.logseries: which should be used for new code. Notes ----- The probability density for the Log Series distribution is .. math:: P(k) = \frac{-p^k}{k \ln(1-p)}, where p = probability. The log series distribution is frequently used to represent species richness and occurrence, first proposed by Fisher, Corbet, and Williams in 1943 [2]. It may also be used to model the numbers of occupants seen in cars [3]. References ---------- .. [1] Buzas, Martin A.; Culver, Stephen J., Understanding regional species diversity through the log series distribution of occurrences: BIODIVERSITY RESEARCH Diversity & Distributions, Volume 5, Number 5, September 1999 , pp. 187-195(9). .. [2] Fisher, R.A,, A.S. Corbet, and C.B. Williams. 1943. The relation between the number of species and the number of individuals in a random sample of an animal population. Journal of Animal Ecology, 12:42-58. .. [3] D. J. Hand, F. Daly, D. Lunn, E. Ostrowski, A Handbook of Small Data Sets, CRC Press, 1994. .. [4] Wikipedia, "Logarithmic distribution", https://en.wikipedia.org/wiki/Logarithmic_distribution Examples -------- Draw samples from the distribution: >>> a = .6 >>> s = np.random.logseries(a, 10000) >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s) # plot against distribution >>> def logseries(k, p): ... return -p**k/(k*np.log(1-p)) >>> plt.plot(bins, logseries(bins, a)*count.max()/ ... logseries(bins, a).max(), 'r') >>> plt.show() hypergeometric(ngood, nbad, nsample, size=None) Draw samples from a Hypergeometric distribution. Samples are drawn from a hypergeometric distribution with specified parameters, `ngood` (ways to make a good selection), `nbad` (ways to make a bad selection), and `nsample` (number of items sampled, which is less than or equal to the sum ``ngood + nbad``). .. note:: New code should use the ``hypergeometric`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- ngood : int or array_like of ints Number of ways to make a good selection. Must be nonnegative. nbad : int or array_like of ints Number of ways to make a bad selection. Must be nonnegative. nsample : int or array_like of ints Number of items sampled. Must be at least 1 and at most ``ngood + nbad``. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if `ngood`, `nbad`, and `nsample` are all scalars. Otherwise, ``np.broadcast(ngood, nbad, nsample).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized hypergeometric distribution. Each sample is the number of good items within a randomly selected subset of size `nsample` taken from a set of `ngood` good items and `nbad` bad items. See Also -------- scipy.stats.hypergeom : probability density function, distribution or cumulative density function, etc. Generator.hypergeometric: which should be used for new code. Notes ----- The probability density for the Hypergeometric distribution is .. math:: P(x) = \frac{\binom{g}{x}\binom{b}{n-x}}{\binom{g+b}{n}}, where :math:`0 \le x \le n` and :math:`n-b \le x \le g` for P(x) the probability of ``x`` good results in the drawn sample, g = `ngood`, b = `nbad`, and n = `nsample`. Consider an urn with black and white marbles in it, `ngood` of them are black and `nbad` are white. If you draw `nsample` balls without replacement, then the hypergeometric distribution describes the distribution of black balls in the drawn sample. Note that this distribution is very similar to the binomial distribution, except that in this case, samples are drawn without replacement, whereas in the Binomial case samples are drawn with replacement (or the sample space is infinite). As the sample space becomes large, this distribution approaches the binomial. References ---------- .. [1] Lentner, Marvin, "Elementary Applied Statistics", Bogden and Quigley, 1972. .. [2] Weisstein, Eric W. "Hypergeometric Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/HypergeometricDistribution.html .. [3] Wikipedia, "Hypergeometric distribution", https://en.wikipedia.org/wiki/Hypergeometric_distribution Examples -------- Draw samples from the distribution: >>> ngood, nbad, nsamp = 100, 2, 10 # number of good, number of bad, and number of samples >>> s = np.random.hypergeometric(ngood, nbad, nsamp, 1000) >>> from matplotlib.pyplot import hist >>> hist(s) # note that it is very unlikely to grab both bad items Suppose you have an urn with 15 white and 15 black marbles. If you pull 15 marbles at random, how likely is it that 12 or more of them are one color? >>> s = np.random.hypergeometric(15, 15, 15, 100000) >>> sum(s>=12)/100000. + sum(s<=3)/100000. # answer = 0.003 ... pretty unlikely! geometric(p, size=None) Draw samples from the geometric distribution. Bernoulli trials are experiments with one of two outcomes: success or failure (an example of such an experiment is flipping a coin). The geometric distribution models the number of trials that must be run in order to achieve success. It is therefore supported on the positive integers, ``k = 1, 2, ...``. The probability mass function of the geometric distribution is .. math:: f(k) = (1 - p)^{k - 1} p where `p` is the probability of success of an individual trial. .. note:: New code should use the ``geometric`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- p : float or array_like of floats The probability of success of an individual trial. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``p`` is a scalar. Otherwise, ``np.array(p).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized geometric distribution. See Also -------- Generator.geometric: which should be used for new code. Examples -------- Draw ten thousand values from the geometric distribution, with the probability of an individual success equal to 0.35: >>> z = np.random.geometric(p=0.35, size=10000) How many trials succeeded after a single run? >>> (z == 1).sum() / 10000. 0.34889999999999999 #random zipf(a, size=None) Draw samples from a Zipf distribution. Samples are drawn from a Zipf distribution with specified parameter `a` > 1. The Zipf distribution (also known as the zeta distribution) is a continuous probability distribution that satisfies Zipf's law: the frequency of an item is inversely proportional to its rank in a frequency table. .. note:: New code should use the ``zipf`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- a : float or array_like of floats Distribution parameter. Must be greater than 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``a`` is a scalar. Otherwise, ``np.array(a).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Zipf distribution. See Also -------- scipy.stats.zipf : probability density function, distribution, or cumulative density function, etc. Generator.zipf: which should be used for new code. Notes ----- The probability density for the Zipf distribution is .. math:: p(x) = \frac{x^{-a}}{\zeta(a)}, where :math:`\zeta` is the Riemann Zeta function. It is named for the American linguist George Kingsley Zipf, who noted that the frequency of any word in a sample of a language is inversely proportional to its rank in the frequency table. References ---------- .. [1] Zipf, G. K., "Selected Studies of the Principle of Relative Frequency in Language," Cambridge, MA: Harvard Univ. Press, 1932. Examples -------- Draw samples from the distribution: >>> a = 2. # parameter >>> s = np.random.zipf(a, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> from scipy import special # doctest: +SKIP Truncate s values at 50 so plot is interesting: >>> count, bins, ignored = plt.hist(s[s<50], 50, density=True) >>> x = np.arange(1., 50.) >>> y = x**(-a) / special.zetac(a) # doctest: +SKIP >>> plt.plot(x, y/max(y), linewidth=2, color='r') # doctest: +SKIP >>> plt.show() poisson(lam=1.0, size=None) Draw samples from a Poisson distribution. The Poisson distribution is the limit of the binomial distribution for large N. .. note:: New code should use the ``poisson`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- lam : float or array_like of floats Expected number of events occurring in a fixed-time interval, must be >= 0. A sequence must be broadcastable over the requested size. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``lam`` is a scalar. Otherwise, ``np.array(lam).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Poisson distribution. See Also -------- Generator.poisson: which should be used for new code. Notes ----- The Poisson distribution .. math:: f(k; \lambda)=\frac{\lambda^k e^{-\lambda}}{k!} For events with an expected separation :math:`\lambda` the Poisson distribution :math:`f(k; \lambda)` describes the probability of :math:`k` events occurring within the observed interval :math:`\lambda`. Because the output is limited to the range of the C int64 type, a ValueError is raised when `lam` is within 10 sigma of the maximum representable value. References ---------- .. [1] Weisstein, Eric W. "Poisson Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/PoissonDistribution.html .. [2] Wikipedia, "Poisson distribution", https://en.wikipedia.org/wiki/Poisson_distribution Examples -------- Draw samples from the distribution: >>> import numpy as np >>> s = np.random.poisson(5, 10000) Display histogram of the sample: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 14, density=True) >>> plt.show() Draw each 100 values for lambda 100 and 500: >>> s = np.random.poisson(lam=(100., 500.), size=(100, 2)) negative_binomial(n, p, size=None) Draw samples from a negative binomial distribution. Samples are drawn from a negative binomial distribution with specified parameters, `n` successes and `p` probability of success where `n` is > 0 and `p` is in the interval [0, 1]. .. note:: New code should use the ``negative_binomial`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- n : float or array_like of floats Parameter of the distribution, > 0. p : float or array_like of floats Parameter of the distribution, >= 0 and <=1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``n`` and ``p`` are both scalars. Otherwise, ``np.broadcast(n, p).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized negative binomial distribution, where each sample is equal to N, the number of failures that occurred before a total of n successes was reached. See Also -------- Generator.negative_binomial: which should be used for new code. Notes ----- The probability mass function of the negative binomial distribution is .. math:: P(N;n,p) = \frac{\Gamma(N+n)}{N!\Gamma(n)}p^{n}(1-p)^{N}, where :math:`n` is the number of successes, :math:`p` is the probability of success, :math:`N+n` is the number of trials, and :math:`\Gamma` is the gamma function. When :math:`n` is an integer, :math:`\frac{\Gamma(N+n)}{N!\Gamma(n)} = \binom{N+n-1}{N}`, which is the more common form of this term in the the pmf. The negative binomial distribution gives the probability of N failures given n successes, with a success on the last trial. If one throws a die repeatedly until the third time a "1" appears, then the probability distribution of the number of non-"1"s that appear before the third "1" is a negative binomial distribution. References ---------- .. [1] Weisstein, Eric W. "Negative Binomial Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/NegativeBinomialDistribution.html .. [2] Wikipedia, "Negative binomial distribution", https://en.wikipedia.org/wiki/Negative_binomial_distribution Examples -------- Draw samples from the distribution: A real world example. A company drills wild-cat oil exploration wells, each with an estimated probability of success of 0.1. What is the probability of having one success for each successive well, that is what is the probability of a single success after drilling 5 wells, after 6 wells, etc.? >>> s = np.random.negative_binomial(1, 0.1, 100000) >>> for i in range(1, 11): # doctest: +SKIP ... probability = sum(s>> n, p = 10, .5 # number of trials, probability of each trial >>> s = np.random.binomial(n, p, 1000) # result of flipping a coin 10 times, tested 1000 times. A real world example. A company drills 9 wild-cat oil exploration wells, each with an estimated probability of success of 0.1. All nine wells fail. What is the probability of that happening? Let's do 20,000 trials of the model, and count the number that generate zero positive results. >>> sum(np.random.binomial(9, 0.1, 20000) == 0)/20000. # answer = 0.38885, or 38%. triangular(left, mode, right, size=None) Draw samples from the triangular distribution over the interval ``[left, right]``. The triangular distribution is a continuous probability distribution with lower limit left, peak at mode, and upper limit right. Unlike the other distributions, these parameters directly define the shape of the pdf. .. note:: New code should use the ``triangular`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- left : float or array_like of floats Lower limit. mode : float or array_like of floats The value where the peak of the distribution occurs. The value must fulfill the condition ``left <= mode <= right``. right : float or array_like of floats Upper limit, must be larger than `left`. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``left``, ``mode``, and ``right`` are all scalars. Otherwise, ``np.broadcast(left, mode, right).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized triangular distribution. See Also -------- Generator.triangular: which should be used for new code. Notes ----- The probability density function for the triangular distribution is .. math:: P(x;l, m, r) = \begin{cases} \frac{2(x-l)}{(r-l)(m-l)}& \text{for $l \leq x \leq m$},\\ \frac{2(r-x)}{(r-l)(r-m)}& \text{for $m \leq x \leq r$},\\ 0& \text{otherwise}. \end{cases} The triangular distribution is often used in ill-defined problems where the underlying distribution is not known, but some knowledge of the limits and mode exists. Often it is used in simulations. References ---------- .. [1] Wikipedia, "Triangular distribution" https://en.wikipedia.org/wiki/Triangular_distribution Examples -------- Draw values from the distribution and plot the histogram: >>> import matplotlib.pyplot as plt >>> h = plt.hist(np.random.triangular(-3, 0, 8, 100000), bins=200, ... density=True) >>> plt.show() wald(mean, scale, size=None) Draw samples from a Wald, or inverse Gaussian, distribution. As the scale approaches infinity, the distribution becomes more like a Gaussian. Some references claim that the Wald is an inverse Gaussian with mean equal to 1, but this is by no means universal. The inverse Gaussian distribution was first studied in relationship to Brownian motion. In 1956 M.C.K. Tweedie used the name inverse Gaussian because there is an inverse relationship between the time to cover a unit distance and distance covered in unit time. .. note:: New code should use the ``wald`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- mean : float or array_like of floats Distribution mean, must be > 0. scale : float or array_like of floats Scale parameter, must be > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``mean`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(mean, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Wald distribution. See Also -------- Generator.wald: which should be used for new code. Notes ----- The probability density function for the Wald distribution is .. math:: P(x;mean,scale) = \sqrt{\frac{scale}{2\pi x^3}}e^ \frac{-scale(x-mean)^2}{2\cdotp mean^2x} As noted above the inverse Gaussian distribution first arise from attempts to model Brownian motion. It is also a competitor to the Weibull for use in reliability modeling and modeling stock returns and interest rate processes. References ---------- .. [1] Brighton Webs Ltd., Wald Distribution, https://web.archive.org/web/20090423014010/http://www.brighton-webs.co.uk:80/distributions/wald.asp .. [2] Chhikara, Raj S., and Folks, J. Leroy, "The Inverse Gaussian Distribution: Theory : Methodology, and Applications", CRC Press, 1988. .. [3] Wikipedia, "Inverse Gaussian distribution" https://en.wikipedia.org/wiki/Inverse_Gaussian_distribution Examples -------- Draw values from the distribution and plot the histogram: >>> import matplotlib.pyplot as plt >>> h = plt.hist(np.random.wald(3, 2, 100000), bins=200, density=True) >>> plt.show() rayleigh(scale=1.0, size=None) Draw samples from a Rayleigh distribution. The :math:`\chi` and Weibull distributions are generalizations of the Rayleigh. .. note:: New code should use the ``rayleigh`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- scale : float or array_like of floats, optional Scale, also equals the mode. Must be non-negative. Default is 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``scale`` is a scalar. Otherwise, ``np.array(scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Rayleigh distribution. See Also -------- Generator.rayleigh: which should be used for new code. Notes ----- The probability density function for the Rayleigh distribution is .. math:: P(x;scale) = \frac{x}{scale^2}e^{\frac{-x^2}{2 \cdotp scale^2}} The Rayleigh distribution would arise, for example, if the East and North components of the wind velocity had identical zero-mean Gaussian distributions. Then the wind speed would have a Rayleigh distribution. References ---------- .. [1] Brighton Webs Ltd., "Rayleigh Distribution," https://web.archive.org/web/20090514091424/http://brighton-webs.co.uk:80/distributions/rayleigh.asp .. [2] Wikipedia, "Rayleigh distribution" https://en.wikipedia.org/wiki/Rayleigh_distribution Examples -------- Draw values from the distribution and plot the histogram >>> from matplotlib.pyplot import hist >>> values = hist(np.random.rayleigh(3, 100000), bins=200, density=True) Wave heights tend to follow a Rayleigh distribution. If the mean wave height is 1 meter, what fraction of waves are likely to be larger than 3 meters? >>> meanvalue = 1 >>> modevalue = np.sqrt(2 / np.pi) * meanvalue >>> s = np.random.rayleigh(modevalue, 1000000) The percentage of waves larger than 3 meters is: >>> 100.*sum(s>3)/1000000. 0.087300000000000003 # random lognormal(mean=0.0, sigma=1.0, size=None) Draw samples from a log-normal distribution. Draw samples from a log-normal distribution with specified mean, standard deviation, and array shape. Note that the mean and standard deviation are not the values for the distribution itself, but of the underlying normal distribution it is derived from. .. note:: New code should use the ``lognormal`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- mean : float or array_like of floats, optional Mean value of the underlying normal distribution. Default is 0. sigma : float or array_like of floats, optional Standard deviation of the underlying normal distribution. Must be non-negative. Default is 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``mean`` and ``sigma`` are both scalars. Otherwise, ``np.broadcast(mean, sigma).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized log-normal distribution. See Also -------- scipy.stats.lognorm : probability density function, distribution, cumulative density function, etc. Generator.lognormal: which should be used for new code. Notes ----- A variable `x` has a log-normal distribution if `log(x)` is normally distributed. The probability density function for the log-normal distribution is: .. math:: p(x) = \frac{1}{\sigma x \sqrt{2\pi}} e^{(-\frac{(ln(x)-\mu)^2}{2\sigma^2})} where :math:`\mu` is the mean and :math:`\sigma` is the standard deviation of the normally distributed logarithm of the variable. A log-normal distribution results if a random variable is the *product* of a large number of independent, identically-distributed variables in the same way that a normal distribution results if the variable is the *sum* of a large number of independent, identically-distributed variables. References ---------- .. [1] Limpert, E., Stahel, W. A., and Abbt, M., "Log-normal Distributions across the Sciences: Keys and Clues," BioScience, Vol. 51, No. 5, May, 2001. https://stat.ethz.ch/~stahel/lognormal/bioscience.pdf .. [2] Reiss, R.D. and Thomas, M., "Statistical Analysis of Extreme Values," Basel: Birkhauser Verlag, 2001, pp. 31-32. Examples -------- Draw samples from the distribution: >>> mu, sigma = 3., 1. # mean and standard deviation >>> s = np.random.lognormal(mu, sigma, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 100, density=True, align='mid') >>> x = np.linspace(min(bins), max(bins), 10000) >>> pdf = (np.exp(-(np.log(x) - mu)**2 / (2 * sigma**2)) ... / (x * sigma * np.sqrt(2 * np.pi))) >>> plt.plot(x, pdf, linewidth=2, color='r') >>> plt.axis('tight') >>> plt.show() Demonstrate that taking the products of random samples from a uniform distribution can be fit well by a log-normal probability density function. >>> # Generate a thousand samples: each is the product of 100 random >>> # values, drawn from a normal distribution. >>> b = [] >>> for i in range(1000): ... a = 10. + np.random.standard_normal(100) ... b.append(np.product(a)) >>> b = np.array(b) / np.min(b) # scale values to be positive >>> count, bins, ignored = plt.hist(b, 100, density=True, align='mid') >>> sigma = np.std(np.log(b)) >>> mu = np.mean(np.log(b)) >>> x = np.linspace(min(bins), max(bins), 10000) >>> pdf = (np.exp(-(np.log(x) - mu)**2 / (2 * sigma**2)) ... / (x * sigma * np.sqrt(2 * np.pi))) >>> plt.plot(x, pdf, color='r', linewidth=2) >>> plt.show() logistic(loc=0.0, scale=1.0, size=None) Draw samples from a logistic distribution. Samples are drawn from a logistic distribution with specified parameters, loc (location or mean, also median), and scale (>0). .. note:: New code should use the ``logistic`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- loc : float or array_like of floats, optional Parameter of the distribution. Default is 0. scale : float or array_like of floats, optional Parameter of the distribution. Must be non-negative. Default is 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``loc`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(loc, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized logistic distribution. See Also -------- scipy.stats.logistic : probability density function, distribution or cumulative density function, etc. Generator.logistic: which should be used for new code. Notes ----- The probability density for the Logistic distribution is .. math:: P(x) = P(x) = \frac{e^{-(x-\mu)/s}}{s(1+e^{-(x-\mu)/s})^2}, where :math:`\mu` = location and :math:`s` = scale. The Logistic distribution is used in Extreme Value problems where it can act as a mixture of Gumbel distributions, in Epidemiology, and by the World Chess Federation (FIDE) where it is used in the Elo ranking system, assuming the performance of each player is a logistically distributed random variable. References ---------- .. [1] Reiss, R.-D. and Thomas M. (2001), "Statistical Analysis of Extreme Values, from Insurance, Finance, Hydrology and Other Fields," Birkhauser Verlag, Basel, pp 132-133. .. [2] Weisstein, Eric W. "Logistic Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/LogisticDistribution.html .. [3] Wikipedia, "Logistic-distribution", https://en.wikipedia.org/wiki/Logistic_distribution Examples -------- Draw samples from the distribution: >>> loc, scale = 10, 1 >>> s = np.random.logistic(loc, scale, 10000) >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, bins=50) # plot against distribution >>> def logist(x, loc, scale): ... return np.exp((loc-x)/scale)/(scale*(1+np.exp((loc-x)/scale))**2) >>> lgst_val = logist(bins, loc, scale) >>> plt.plot(bins, lgst_val * count.max() / lgst_val.max()) >>> plt.show() gumbel(loc=0.0, scale=1.0, size=None) Draw samples from a Gumbel distribution. Draw samples from a Gumbel distribution with specified location and scale. For more information on the Gumbel distribution, see Notes and References below. .. note:: New code should use the ``gumbel`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- loc : float or array_like of floats, optional The location of the mode of the distribution. Default is 0. scale : float or array_like of floats, optional The scale parameter of the distribution. Default is 1. Must be non- negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``loc`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(loc, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Gumbel distribution. See Also -------- scipy.stats.gumbel_l scipy.stats.gumbel_r scipy.stats.genextreme weibull Generator.gumbel: which should be used for new code. Notes ----- The Gumbel (or Smallest Extreme Value (SEV) or the Smallest Extreme Value Type I) distribution is one of a class of Generalized Extreme Value (GEV) distributions used in modeling extreme value problems. The Gumbel is a special case of the Extreme Value Type I distribution for maximums from distributions with "exponential-like" tails. The probability density for the Gumbel distribution is .. math:: p(x) = \frac{e^{-(x - \mu)/ \beta}}{\beta} e^{ -e^{-(x - \mu)/ \beta}}, where :math:`\mu` is the mode, a location parameter, and :math:`\beta` is the scale parameter. The Gumbel (named for German mathematician Emil Julius Gumbel) was used very early in the hydrology literature, for modeling the occurrence of flood events. It is also used for modeling maximum wind speed and rainfall rates. It is a "fat-tailed" distribution - the probability of an event in the tail of the distribution is larger than if one used a Gaussian, hence the surprisingly frequent occurrence of 100-year floods. Floods were initially modeled as a Gaussian process, which underestimated the frequency of extreme events. It is one of a class of extreme value distributions, the Generalized Extreme Value (GEV) distributions, which also includes the Weibull and Frechet. The function has a mean of :math:`\mu + 0.57721\beta` and a variance of :math:`\frac{\pi^2}{6}\beta^2`. References ---------- .. [1] Gumbel, E. J., "Statistics of Extremes," New York: Columbia University Press, 1958. .. [2] Reiss, R.-D. and Thomas, M., "Statistical Analysis of Extreme Values from Insurance, Finance, Hydrology and Other Fields," Basel: Birkhauser Verlag, 2001. Examples -------- Draw samples from the distribution: >>> mu, beta = 0, 0.1 # location and scale >>> s = np.random.gumbel(mu, beta, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, density=True) >>> plt.plot(bins, (1/beta)*np.exp(-(bins - mu)/beta) ... * np.exp( -np.exp( -(bins - mu) /beta) ), ... linewidth=2, color='r') >>> plt.show() Show how an extreme value distribution can arise from a Gaussian process and compare to a Gaussian: >>> means = [] >>> maxima = [] >>> for i in range(0,1000) : ... a = np.random.normal(mu, beta, 1000) ... means.append(a.mean()) ... maxima.append(a.max()) >>> count, bins, ignored = plt.hist(maxima, 30, density=True) >>> beta = np.std(maxima) * np.sqrt(6) / np.pi >>> mu = np.mean(maxima) - 0.57721*beta >>> plt.plot(bins, (1/beta)*np.exp(-(bins - mu)/beta) ... * np.exp(-np.exp(-(bins - mu)/beta)), ... linewidth=2, color='r') >>> plt.plot(bins, 1/(beta * np.sqrt(2 * np.pi)) ... * np.exp(-(bins - mu)**2 / (2 * beta**2)), ... linewidth=2, color='g') >>> plt.show() laplace(loc=0.0, scale=1.0, size=None) Draw samples from the Laplace or double exponential distribution with specified location (or mean) and scale (decay). The Laplace distribution is similar to the Gaussian/normal distribution, but is sharper at the peak and has fatter tails. It represents the difference between two independent, identically distributed exponential random variables. .. note:: New code should use the ``laplace`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- loc : float or array_like of floats, optional The position, :math:`\mu`, of the distribution peak. Default is 0. scale : float or array_like of floats, optional :math:`\lambda`, the exponential decay. Default is 1. Must be non- negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``loc`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(loc, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Laplace distribution. See Also -------- Generator.laplace: which should be used for new code. Notes ----- It has the probability density function .. math:: f(x; \mu, \lambda) = \frac{1}{2\lambda} \exp\left(-\frac{|x - \mu|}{\lambda}\right). The first law of Laplace, from 1774, states that the frequency of an error can be expressed as an exponential function of the absolute magnitude of the error, which leads to the Laplace distribution. For many problems in economics and health sciences, this distribution seems to model the data better than the standard Gaussian distribution. References ---------- .. [1] Abramowitz, M. and Stegun, I. A. (Eds.). "Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing," New York: Dover, 1972. .. [2] Kotz, Samuel, et. al. "The Laplace Distribution and Generalizations, " Birkhauser, 2001. .. [3] Weisstein, Eric W. "Laplace Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/LaplaceDistribution.html .. [4] Wikipedia, "Laplace distribution", https://en.wikipedia.org/wiki/Laplace_distribution Examples -------- Draw samples from the distribution >>> loc, scale = 0., 1. >>> s = np.random.laplace(loc, scale, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, density=True) >>> x = np.arange(-8., 8., .01) >>> pdf = np.exp(-abs(x-loc)/scale)/(2.*scale) >>> plt.plot(x, pdf) Plot Gaussian for comparison: >>> g = (1/(scale * np.sqrt(2 * np.pi)) * ... np.exp(-(x - loc)**2 / (2 * scale**2))) >>> plt.plot(x,g) power(a, size=None) Draws samples in [0, 1] from a power distribution with positive exponent a - 1. Also known as the power function distribution. .. note:: New code should use the ``power`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- a : float or array_like of floats Parameter of the distribution. Must be non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``a`` is a scalar. Otherwise, ``np.array(a).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized power distribution. Raises ------ ValueError If a < 1. See Also -------- Generator.power: which should be used for new code. Notes ----- The probability density function is .. math:: P(x; a) = ax^{a-1}, 0 \le x \le 1, a>0. The power function distribution is just the inverse of the Pareto distribution. It may also be seen as a special case of the Beta distribution. It is used, for example, in modeling the over-reporting of insurance claims. References ---------- .. [1] Christian Kleiber, Samuel Kotz, "Statistical size distributions in economics and actuarial sciences", Wiley, 2003. .. [2] Heckert, N. A. and Filliben, James J. "NIST Handbook 148: Dataplot Reference Manual, Volume 2: Let Subcommands and Library Functions", National Institute of Standards and Technology Handbook Series, June 2003. https://www.itl.nist.gov/div898/software/dataplot/refman2/auxillar/powpdf.pdf Examples -------- Draw samples from the distribution: >>> a = 5. # shape >>> samples = 1000 >>> s = np.random.power(a, samples) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, bins=30) >>> x = np.linspace(0, 1, 100) >>> y = a*x**(a-1.) >>> normed_y = samples*np.diff(bins)[0]*y >>> plt.plot(x, normed_y) >>> plt.show() Compare the power function distribution to the inverse of the Pareto. >>> from scipy import stats # doctest: +SKIP >>> rvs = np.random.power(5, 1000000) >>> rvsp = np.random.pareto(5, 1000000) >>> xx = np.linspace(0,1,100) >>> powpdf = stats.powerlaw.pdf(xx,5) # doctest: +SKIP >>> plt.figure() >>> plt.hist(rvs, bins=50, density=True) >>> plt.plot(xx,powpdf,'r-') # doctest: +SKIP >>> plt.title('np.random.power(5)') >>> plt.figure() >>> plt.hist(1./(1.+rvsp), bins=50, density=True) >>> plt.plot(xx,powpdf,'r-') # doctest: +SKIP >>> plt.title('inverse of 1 + np.random.pareto(5)') >>> plt.figure() >>> plt.hist(1./(1.+rvsp), bins=50, density=True) >>> plt.plot(xx,powpdf,'r-') # doctest: +SKIP >>> plt.title('inverse of stats.pareto(5)') weibull(a, size=None) Draw samples from a Weibull distribution. Draw samples from a 1-parameter Weibull distribution with the given shape parameter `a`. .. math:: X = (-ln(U))^{1/a} Here, U is drawn from the uniform distribution over (0,1]. The more common 2-parameter Weibull, including a scale parameter :math:`\lambda` is just :math:`X = \lambda(-ln(U))^{1/a}`. .. note:: New code should use the ``weibull`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- a : float or array_like of floats Shape parameter of the distribution. Must be nonnegative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``a`` is a scalar. Otherwise, ``np.array(a).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Weibull distribution. See Also -------- scipy.stats.weibull_max scipy.stats.weibull_min scipy.stats.genextreme gumbel Generator.weibull: which should be used for new code. Notes ----- The Weibull (or Type III asymptotic extreme value distribution for smallest values, SEV Type III, or Rosin-Rammler distribution) is one of a class of Generalized Extreme Value (GEV) distributions used in modeling extreme value problems. This class includes the Gumbel and Frechet distributions. The probability density for the Weibull distribution is .. math:: p(x) = \frac{a} {\lambda}(\frac{x}{\lambda})^{a-1}e^{-(x/\lambda)^a}, where :math:`a` is the shape and :math:`\lambda` the scale. The function has its peak (the mode) at :math:`\lambda(\frac{a-1}{a})^{1/a}`. When ``a = 1``, the Weibull distribution reduces to the exponential distribution. References ---------- .. [1] Waloddi Weibull, Royal Technical University, Stockholm, 1939 "A Statistical Theory Of The Strength Of Materials", Ingeniorsvetenskapsakademiens Handlingar Nr 151, 1939, Generalstabens Litografiska Anstalts Forlag, Stockholm. .. [2] Waloddi Weibull, "A Statistical Distribution Function of Wide Applicability", Journal Of Applied Mechanics ASME Paper 1951. .. [3] Wikipedia, "Weibull distribution", https://en.wikipedia.org/wiki/Weibull_distribution Examples -------- Draw samples from the distribution: >>> a = 5. # shape >>> s = np.random.weibull(a, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> x = np.arange(1,100.)/50. >>> def weib(x,n,a): ... return (a / n) * (x / n)**(a - 1) * np.exp(-(x / n)**a) >>> count, bins, ignored = plt.hist(np.random.weibull(5.,1000)) >>> x = np.arange(1,100.)/50. >>> scale = count.max()/weib(x, 1., 5.).max() >>> plt.plot(x, weib(x, 1., 5.)*scale) >>> plt.show() pareto(a, size=None) Draw samples from a Pareto II or Lomax distribution with specified shape. The Lomax or Pareto II distribution is a shifted Pareto distribution. The classical Pareto distribution can be obtained from the Lomax distribution by adding 1 and multiplying by the scale parameter ``m`` (see Notes). The smallest value of the Lomax distribution is zero while for the classical Pareto distribution it is ``mu``, where the standard Pareto distribution has location ``mu = 1``. Lomax can also be considered as a simplified version of the Generalized Pareto distribution (available in SciPy), with the scale set to one and the location set to zero. The Pareto distribution must be greater than zero, and is unbounded above. It is also known as the "80-20 rule". In this distribution, 80 percent of the weights are in the lowest 20 percent of the range, while the other 20 percent fill the remaining 80 percent of the range. .. note:: New code should use the ``pareto`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- a : float or array_like of floats Shape of the distribution. Must be positive. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``a`` is a scalar. Otherwise, ``np.array(a).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Pareto distribution. See Also -------- scipy.stats.lomax : probability density function, distribution or cumulative density function, etc. scipy.stats.genpareto : probability density function, distribution or cumulative density function, etc. Generator.pareto: which should be used for new code. Notes ----- The probability density for the Pareto distribution is .. math:: p(x) = \frac{am^a}{x^{a+1}} where :math:`a` is the shape and :math:`m` the scale. The Pareto distribution, named after the Italian economist Vilfredo Pareto, is a power law probability distribution useful in many real world problems. Outside the field of economics it is generally referred to as the Bradford distribution. Pareto developed the distribution to describe the distribution of wealth in an economy. It has also found use in insurance, web page access statistics, oil field sizes, and many other problems, including the download frequency for projects in Sourceforge [1]_. It is one of the so-called "fat-tailed" distributions. References ---------- .. [1] Francis Hunt and Paul Johnson, On the Pareto Distribution of Sourceforge projects. .. [2] Pareto, V. (1896). Course of Political Economy. Lausanne. .. [3] Reiss, R.D., Thomas, M.(2001), Statistical Analysis of Extreme Values, Birkhauser Verlag, Basel, pp 23-30. .. [4] Wikipedia, "Pareto distribution", https://en.wikipedia.org/wiki/Pareto_distribution Examples -------- Draw samples from the distribution: >>> a, m = 3., 2. # shape and mode >>> s = (np.random.pareto(a, 1000) + 1) * m Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, _ = plt.hist(s, 100, density=True) >>> fit = a*m**a / bins**(a+1) >>> plt.plot(bins, max(count)*fit/max(fit), linewidth=2, color='r') >>> plt.show() vonmises(mu, kappa, size=None) Draw samples from a von Mises distribution. Samples are drawn from a von Mises distribution with specified mode (mu) and dispersion (kappa), on the interval [-pi, pi]. The von Mises distribution (also known as the circular normal distribution) is a continuous probability distribution on the unit circle. It may be thought of as the circular analogue of the normal distribution. .. note:: New code should use the ``vonmises`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- mu : float or array_like of floats Mode ("center") of the distribution. kappa : float or array_like of floats Dispersion of the distribution, has to be >=0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``mu`` and ``kappa`` are both scalars. Otherwise, ``np.broadcast(mu, kappa).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized von Mises distribution. See Also -------- scipy.stats.vonmises : probability density function, distribution, or cumulative density function, etc. Generator.vonmises: which should be used for new code. Notes ----- The probability density for the von Mises distribution is .. math:: p(x) = \frac{e^{\kappa cos(x-\mu)}}{2\pi I_0(\kappa)}, where :math:`\mu` is the mode and :math:`\kappa` the dispersion, and :math:`I_0(\kappa)` is the modified Bessel function of order 0. The von Mises is named for Richard Edler von Mises, who was born in Austria-Hungary, in what is now the Ukraine. He fled to the United States in 1939 and became a professor at Harvard. He worked in probability theory, aerodynamics, fluid mechanics, and philosophy of science. References ---------- .. [1] Abramowitz, M. and Stegun, I. A. (Eds.). "Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing," New York: Dover, 1972. .. [2] von Mises, R., "Mathematical Theory of Probability and Statistics", New York: Academic Press, 1964. Examples -------- Draw samples from the distribution: >>> mu, kappa = 0.0, 4.0 # mean and dispersion >>> s = np.random.vonmises(mu, kappa, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> from scipy.special import i0 # doctest: +SKIP >>> plt.hist(s, 50, density=True) >>> x = np.linspace(-np.pi, np.pi, num=51) >>> y = np.exp(kappa*np.cos(x-mu))/(2*np.pi*i0(kappa)) # doctest: +SKIP >>> plt.plot(x, y, linewidth=2, color='r') # doctest: +SKIP >>> plt.show() standard_t(df, size=None) Draw samples from a standard Student's t distribution with `df` degrees of freedom. A special case of the hyperbolic distribution. As `df` gets large, the result resembles that of the standard normal distribution (`standard_normal`). .. note:: New code should use the ``standard_t`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- df : float or array_like of floats Degrees of freedom, must be > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``df`` is a scalar. Otherwise, ``np.array(df).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized standard Student's t distribution. See Also -------- Generator.standard_t: which should be used for new code. Notes ----- The probability density function for the t distribution is .. math:: P(x, df) = \frac{\Gamma(\frac{df+1}{2})}{\sqrt{\pi df} \Gamma(\frac{df}{2})}\Bigl( 1+\frac{x^2}{df} \Bigr)^{-(df+1)/2} The t test is based on an assumption that the data come from a Normal distribution. The t test provides a way to test whether the sample mean (that is the mean calculated from the data) is a good estimate of the true mean. The derivation of the t-distribution was first published in 1908 by William Gosset while working for the Guinness Brewery in Dublin. Due to proprietary issues, he had to publish under a pseudonym, and so he used the name Student. References ---------- .. [1] Dalgaard, Peter, "Introductory Statistics With R", Springer, 2002. .. [2] Wikipedia, "Student's t-distribution" https://en.wikipedia.org/wiki/Student's_t-distribution Examples -------- From Dalgaard page 83 [1]_, suppose the daily energy intake for 11 women in kilojoules (kJ) is: >>> intake = np.array([5260., 5470, 5640, 6180, 6390, 6515, 6805, 7515, \ ... 7515, 8230, 8770]) Does their energy intake deviate systematically from the recommended value of 7725 kJ? Our null hypothesis will be the absence of deviation, and the alternate hypothesis will be the presence of an effect that could be either positive or negative, hence making our test 2-tailed. Because we are estimating the mean and we have N=11 values in our sample, we have N-1=10 degrees of freedom. We set our significance level to 95% and compute the t statistic using the empirical mean and empirical standard deviation of our intake. We use a ddof of 1 to base the computation of our empirical standard deviation on an unbiased estimate of the variance (note: the final estimate is not unbiased due to the concave nature of the square root). >>> np.mean(intake) 6753.636363636364 >>> intake.std(ddof=1) 1142.1232221373727 >>> t = (np.mean(intake)-7725)/(intake.std(ddof=1)/np.sqrt(len(intake))) >>> t -2.8207540608310198 We draw 1000000 samples from Student's t distribution with the adequate degrees of freedom. >>> import matplotlib.pyplot as plt >>> s = np.random.standard_t(10, size=1000000) >>> h = plt.hist(s, bins=100, density=True) Does our t statistic land in one of the two critical regions found at both tails of the distribution? >>> np.sum(np.abs(t) < np.abs(s)) / float(len(s)) 0.018318 #random < 0.05, statistic is in critical region The probability value for this 2-tailed test is about 1.83%, which is lower than the 5% pre-determined significance threshold. Therefore, the probability of observing values as extreme as our intake conditionally on the null hypothesis being true is too low, and we reject the null hypothesis of no deviation. standard_cauchy(size=None) Draw samples from a standard Cauchy distribution with mode = 0. Also known as the Lorentz distribution. .. note:: New code should use the ``standard_cauchy`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- samples : ndarray or scalar The drawn samples. See Also -------- Generator.standard_cauchy: which should be used for new code. Notes ----- The probability density function for the full Cauchy distribution is .. math:: P(x; x_0, \gamma) = \frac{1}{\pi \gamma \bigl[ 1+ (\frac{x-x_0}{\gamma})^2 \bigr] } and the Standard Cauchy distribution just sets :math:`x_0=0` and :math:`\gamma=1` The Cauchy distribution arises in the solution to the driven harmonic oscillator problem, and also describes spectral line broadening. It also describes the distribution of values at which a line tilted at a random angle will cut the x axis. When studying hypothesis tests that assume normality, seeing how the tests perform on data from a Cauchy distribution is a good indicator of their sensitivity to a heavy-tailed distribution, since the Cauchy looks very much like a Gaussian distribution, but with heavier tails. References ---------- .. [1] NIST/SEMATECH e-Handbook of Statistical Methods, "Cauchy Distribution", https://www.itl.nist.gov/div898/handbook/eda/section3/eda3663.htm .. [2] Weisstein, Eric W. "Cauchy Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/CauchyDistribution.html .. [3] Wikipedia, "Cauchy distribution" https://en.wikipedia.org/wiki/Cauchy_distribution Examples -------- Draw samples and plot the distribution: >>> import matplotlib.pyplot as plt >>> s = np.random.standard_cauchy(1000000) >>> s = s[(s>-25) & (s<25)] # truncate distribution so it plots well >>> plt.hist(s, bins=100) >>> plt.show() noncentral_chisquare(df, nonc, size=None) Draw samples from a noncentral chi-square distribution. The noncentral :math:`\chi^2` distribution is a generalization of the :math:`\chi^2` distribution. .. note:: New code should use the ``noncentral_chisquare`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- df : float or array_like of floats Degrees of freedom, must be > 0. .. versionchanged:: 1.10.0 Earlier NumPy versions required dfnum > 1. nonc : float or array_like of floats Non-centrality, must be non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``df`` and ``nonc`` are both scalars. Otherwise, ``np.broadcast(df, nonc).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized noncentral chi-square distribution. See Also -------- Generator.noncentral_chisquare: which should be used for new code. Notes ----- The probability density function for the noncentral Chi-square distribution is .. math:: P(x;df,nonc) = \sum^{\infty}_{i=0} \frac{e^{-nonc/2}(nonc/2)^{i}}{i!} P_{Y_{df+2i}}(x), where :math:`Y_{q}` is the Chi-square with q degrees of freedom. References ---------- .. [1] Wikipedia, "Noncentral chi-squared distribution" https://en.wikipedia.org/wiki/Noncentral_chi-squared_distribution Examples -------- Draw values from the distribution and plot the histogram >>> import matplotlib.pyplot as plt >>> values = plt.hist(np.random.noncentral_chisquare(3, 20, 100000), ... bins=200, density=True) >>> plt.show() Draw values from a noncentral chisquare with very small noncentrality, and compare to a chisquare. >>> plt.figure() >>> values = plt.hist(np.random.noncentral_chisquare(3, .0000001, 100000), ... bins=np.arange(0., 25, .1), density=True) >>> values2 = plt.hist(np.random.chisquare(3, 100000), ... bins=np.arange(0., 25, .1), density=True) >>> plt.plot(values[1][0:-1], values[0]-values2[0], 'ob') >>> plt.show() Demonstrate how large values of non-centrality lead to a more symmetric distribution. >>> plt.figure() >>> values = plt.hist(np.random.noncentral_chisquare(3, 20, 100000), ... bins=200, density=True) >>> plt.show() chisquare(df, size=None) Draw samples from a chi-square distribution. When `df` independent random variables, each with standard normal distributions (mean 0, variance 1), are squared and summed, the resulting distribution is chi-square (see Notes). This distribution is often used in hypothesis testing. .. note:: New code should use the ``chisquare`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- df : float or array_like of floats Number of degrees of freedom, must be > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``df`` is a scalar. Otherwise, ``np.array(df).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized chi-square distribution. Raises ------ ValueError When `df` <= 0 or when an inappropriate `size` (e.g. ``size=-1``) is given. See Also -------- Generator.chisquare: which should be used for new code. Notes ----- The variable obtained by summing the squares of `df` independent, standard normally distributed random variables: .. math:: Q = \sum_{i=0}^{\mathtt{df}} X^2_i is chi-square distributed, denoted .. math:: Q \sim \chi^2_k. The probability density function of the chi-squared distribution is .. math:: p(x) = \frac{(1/2)^{k/2}}{\Gamma(k/2)} x^{k/2 - 1} e^{-x/2}, where :math:`\Gamma` is the gamma function, .. math:: \Gamma(x) = \int_0^{-\infty} t^{x - 1} e^{-t} dt. References ---------- .. [1] NIST "Engineering Statistics Handbook" https://www.itl.nist.gov/div898/handbook/eda/section3/eda3666.htm Examples -------- >>> np.random.chisquare(2,4) array([ 1.89920014, 9.00867716, 3.13710533, 5.62318272]) # random noncentral_f(dfnum, dfden, nonc, size=None) Draw samples from the noncentral F distribution. Samples are drawn from an F distribution with specified parameters, `dfnum` (degrees of freedom in numerator) and `dfden` (degrees of freedom in denominator), where both parameters > 1. `nonc` is the non-centrality parameter. .. note:: New code should use the ``noncentral_f`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- dfnum : float or array_like of floats Numerator degrees of freedom, must be > 0. .. versionchanged:: 1.14.0 Earlier NumPy versions required dfnum > 1. dfden : float or array_like of floats Denominator degrees of freedom, must be > 0. nonc : float or array_like of floats Non-centrality parameter, the sum of the squares of the numerator means, must be >= 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``dfnum``, ``dfden``, and ``nonc`` are all scalars. Otherwise, ``np.broadcast(dfnum, dfden, nonc).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized noncentral Fisher distribution. See Also -------- Generator.noncentral_f: which should be used for new code. Notes ----- When calculating the power of an experiment (power = probability of rejecting the null hypothesis when a specific alternative is true) the non-central F statistic becomes important. When the null hypothesis is true, the F statistic follows a central F distribution. When the null hypothesis is not true, then it follows a non-central F statistic. References ---------- .. [1] Weisstein, Eric W. "Noncentral F-Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/NoncentralF-Distribution.html .. [2] Wikipedia, "Noncentral F-distribution", https://en.wikipedia.org/wiki/Noncentral_F-distribution Examples -------- In a study, testing for a specific alternative to the null hypothesis requires use of the Noncentral F distribution. We need to calculate the area in the tail of the distribution that exceeds the value of the F distribution for the null hypothesis. We'll plot the two probability distributions for comparison. >>> dfnum = 3 # between group deg of freedom >>> dfden = 20 # within groups degrees of freedom >>> nonc = 3.0 >>> nc_vals = np.random.noncentral_f(dfnum, dfden, nonc, 1000000) >>> NF = np.histogram(nc_vals, bins=50, density=True) >>> c_vals = np.random.f(dfnum, dfden, 1000000) >>> F = np.histogram(c_vals, bins=50, density=True) >>> import matplotlib.pyplot as plt >>> plt.plot(F[1][1:], F[0]) >>> plt.plot(NF[1][1:], NF[0]) >>> plt.show() f(dfnum, dfden, size=None) Draw samples from an F distribution. Samples are drawn from an F distribution with specified parameters, `dfnum` (degrees of freedom in numerator) and `dfden` (degrees of freedom in denominator), where both parameters must be greater than zero. The random variate of the F distribution (also known as the Fisher distribution) is a continuous probability distribution that arises in ANOVA tests, and is the ratio of two chi-square variates. .. note:: New code should use the ``f`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- dfnum : float or array_like of floats Degrees of freedom in numerator, must be > 0. dfden : float or array_like of float Degrees of freedom in denominator, must be > 0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``dfnum`` and ``dfden`` are both scalars. Otherwise, ``np.broadcast(dfnum, dfden).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized Fisher distribution. See Also -------- scipy.stats.f : probability density function, distribution or cumulative density function, etc. Generator.f: which should be used for new code. Notes ----- The F statistic is used to compare in-group variances to between-group variances. Calculating the distribution depends on the sampling, and so it is a function of the respective degrees of freedom in the problem. The variable `dfnum` is the number of samples minus one, the between-groups degrees of freedom, while `dfden` is the within-groups degrees of freedom, the sum of the number of samples in each group minus the number of groups. References ---------- .. [1] Glantz, Stanton A. "Primer of Biostatistics.", McGraw-Hill, Fifth Edition, 2002. .. [2] Wikipedia, "F-distribution", https://en.wikipedia.org/wiki/F-distribution Examples -------- An example from Glantz[1], pp 47-40: Two groups, children of diabetics (25 people) and children from people without diabetes (25 controls). Fasting blood glucose was measured, case group had a mean value of 86.1, controls had a mean value of 82.2. Standard deviations were 2.09 and 2.49 respectively. Are these data consistent with the null hypothesis that the parents diabetic status does not affect their children's blood glucose levels? Calculating the F statistic from the data gives a value of 36.01. Draw samples from the distribution: >>> dfnum = 1. # between group degrees of freedom >>> dfden = 48. # within groups degrees of freedom >>> s = np.random.f(dfnum, dfden, 1000) The lower bound for the top 1% of the samples is : >>> np.sort(s)[-10] 7.61988120985 # random So there is about a 1% chance that the F statistic will exceed 7.62, the measured value is 36, so the null hypothesis is rejected at the 1% level. gamma(shape, scale=1.0, size=None) Draw samples from a Gamma distribution. Samples are drawn from a Gamma distribution with specified parameters, `shape` (sometimes designated "k") and `scale` (sometimes designated "theta"), where both parameters are > 0. .. note:: New code should use the ``gamma`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- shape : float or array_like of floats The shape of the gamma distribution. Must be non-negative. scale : float or array_like of floats, optional The scale of the gamma distribution. Must be non-negative. Default is equal to 1. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``shape`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(shape, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized gamma distribution. See Also -------- scipy.stats.gamma : probability density function, distribution or cumulative density function, etc. Generator.gamma: which should be used for new code. Notes ----- The probability density for the Gamma distribution is .. math:: p(x) = x^{k-1}\frac{e^{-x/\theta}}{\theta^k\Gamma(k)}, where :math:`k` is the shape and :math:`\theta` the scale, and :math:`\Gamma` is the Gamma function. The Gamma distribution is often used to model the times to failure of electronic components, and arises naturally in processes for which the waiting times between Poisson distributed events are relevant. References ---------- .. [1] Weisstein, Eric W. "Gamma Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/GammaDistribution.html .. [2] Wikipedia, "Gamma distribution", https://en.wikipedia.org/wiki/Gamma_distribution Examples -------- Draw samples from the distribution: >>> shape, scale = 2., 2. # mean=4, std=2*sqrt(2) >>> s = np.random.gamma(shape, scale, 1000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> import scipy.special as sps # doctest: +SKIP >>> count, bins, ignored = plt.hist(s, 50, density=True) >>> y = bins**(shape-1)*(np.exp(-bins/scale) / # doctest: +SKIP ... (sps.gamma(shape)*scale**shape)) >>> plt.plot(bins, y, linewidth=2, color='r') # doctest: +SKIP >>> plt.show() standard_gamma(shape, size=None) Draw samples from a standard Gamma distribution. Samples are drawn from a Gamma distribution with specified parameters, shape (sometimes designated "k") and scale=1. .. note:: New code should use the ``standard_gamma`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- shape : float or array_like of floats Parameter, must be non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``shape`` is a scalar. Otherwise, ``np.array(shape).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized standard gamma distribution. See Also -------- scipy.stats.gamma : probability density function, distribution or cumulative density function, etc. Generator.standard_gamma: which should be used for new code. Notes ----- The probability density for the Gamma distribution is .. math:: p(x) = x^{k-1}\frac{e^{-x/\theta}}{\theta^k\Gamma(k)}, where :math:`k` is the shape and :math:`\theta` the scale, and :math:`\Gamma` is the Gamma function. The Gamma distribution is often used to model the times to failure of electronic components, and arises naturally in processes for which the waiting times between Poisson distributed events are relevant. References ---------- .. [1] Weisstein, Eric W. "Gamma Distribution." From MathWorld--A Wolfram Web Resource. http://mathworld.wolfram.com/GammaDistribution.html .. [2] Wikipedia, "Gamma distribution", https://en.wikipedia.org/wiki/Gamma_distribution Examples -------- Draw samples from the distribution: >>> shape, scale = 2., 1. # mean and width >>> s = np.random.standard_gamma(shape, 1000000) Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> import scipy.special as sps # doctest: +SKIP >>> count, bins, ignored = plt.hist(s, 50, density=True) >>> y = bins**(shape-1) * ((np.exp(-bins/scale))/ # doctest: +SKIP ... (sps.gamma(shape) * scale**shape)) >>> plt.plot(bins, y, linewidth=2, color='r') # doctest: +SKIP >>> plt.show() normal(loc=0.0, scale=1.0, size=None) Draw random samples from a normal (Gaussian) distribution. The probability density function of the normal distribution, first derived by De Moivre and 200 years later by both Gauss and Laplace independently [2]_, is often called the bell curve because of its characteristic shape (see the example below). The normal distributions occurs often in nature. For example, it describes the commonly occurring distribution of samples influenced by a large number of tiny, random disturbances, each with its own unique distribution [2]_. .. note:: New code should use the ``normal`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- loc : float or array_like of floats Mean ("centre") of the distribution. scale : float or array_like of floats Standard deviation (spread or "width") of the distribution. Must be non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``loc`` and ``scale`` are both scalars. Otherwise, ``np.broadcast(loc, scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized normal distribution. See Also -------- scipy.stats.norm : probability density function, distribution or cumulative density function, etc. Generator.normal: which should be used for new code. Notes ----- The probability density for the Gaussian distribution is .. math:: p(x) = \frac{1}{\sqrt{ 2 \pi \sigma^2 }} e^{ - \frac{ (x - \mu)^2 } {2 \sigma^2} }, where :math:`\mu` is the mean and :math:`\sigma` the standard deviation. The square of the standard deviation, :math:`\sigma^2`, is called the variance. The function has its peak at the mean, and its "spread" increases with the standard deviation (the function reaches 0.607 times its maximum at :math:`x + \sigma` and :math:`x - \sigma` [2]_). This implies that normal is more likely to return samples lying close to the mean, rather than those far away. References ---------- .. [1] Wikipedia, "Normal distribution", https://en.wikipedia.org/wiki/Normal_distribution .. [2] P. R. Peebles Jr., "Central Limit Theorem" in "Probability, Random Variables and Random Signal Principles", 4th ed., 2001, pp. 51, 51, 125. Examples -------- Draw samples from the distribution: >>> mu, sigma = 0, 0.1 # mean and standard deviation >>> s = np.random.normal(mu, sigma, 1000) Verify the mean and the variance: >>> abs(mu - np.mean(s)) 0.0 # may vary >>> abs(sigma - np.std(s, ddof=1)) 0.1 # may vary Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 30, density=True) >>> plt.plot(bins, 1/(sigma * np.sqrt(2 * np.pi)) * ... np.exp( - (bins - mu)**2 / (2 * sigma**2) ), ... linewidth=2, color='r') >>> plt.show() Two-by-four array of samples from N(3, 6.25): >>> np.random.normal(3, 2.5, size=(2, 4)) array([[-4.49401501, 4.00950034, -1.81814867, 7.29718677], # random [ 0.39924804, 4.68456316, 4.99394529, 4.84057254]]) # random standard_normal(size=None) Draw samples from a standard Normal distribution (mean=0, stdev=1). .. note:: New code should use the ``standard_normal`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray A floating-point array of shape ``size`` of drawn samples, or a single sample if ``size`` was not specified. See Also -------- normal : Equivalent function with additional ``loc`` and ``scale`` arguments for setting the mean and standard deviation. Generator.standard_normal: which should be used for new code. Notes ----- For random samples from :math:`N(\mu, \sigma^2)`, use one of:: mu + sigma * np.random.standard_normal(size=...) np.random.normal(mu, sigma, size=...) Examples -------- >>> np.random.standard_normal() 2.1923875335537315 #random >>> s = np.random.standard_normal(8000) >>> s array([ 0.6888893 , 0.78096262, -0.89086505, ..., 0.49876311, # random -0.38672696, -0.4685006 ]) # random >>> s.shape (8000,) >>> s = np.random.standard_normal(size=(3, 4, 2)) >>> s.shape (3, 4, 2) Two-by-four array of samples from :math:`N(3, 6.25)`: >>> 3 + 2.5 * np.random.standard_normal(size=(2, 4)) array([[-4.49401501, 4.00950034, -1.81814867, 7.29718677], # random [ 0.39924804, 4.68456316, 4.99394529, 4.84057254]]) # random random_integers(low, high=None, size=None) Random integers of type `np.int_` between `low` and `high`, inclusive. Return random integers of type `np.int_` from the "discrete uniform" distribution in the closed interval [`low`, `high`]. If `high` is None (the default), then results are from [1, `low`]. The `np.int_` type translates to the C long integer type and its precision is platform dependent. This function has been deprecated. Use randint instead. .. deprecated:: 1.11.0 Parameters ---------- low : int Lowest (signed) integer to be drawn from the distribution (unless ``high=None``, in which case this parameter is the *highest* such integer). high : int, optional If provided, the largest (signed) integer to be drawn from the distribution (see above for behavior if ``high=None``). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : int or ndarray of ints `size`-shaped array of random integers from the appropriate distribution, or a single such random int if `size` not provided. See Also -------- randint : Similar to `random_integers`, only for the half-open interval [`low`, `high`), and 0 is the lowest value if `high` is omitted. Notes ----- To sample from N evenly spaced floating-point numbers between a and b, use:: a + (b - a) * (np.random.random_integers(N) - 1) / (N - 1.) Examples -------- >>> np.random.random_integers(5) 4 # random >>> type(np.random.random_integers(5)) >>> np.random.random_integers(5, size=(3,2)) array([[5, 4], # random [3, 3], [4, 5]]) Choose five random numbers from the set of five evenly-spaced numbers between 0 and 2.5, inclusive (*i.e.*, from the set :math:`{0, 5/8, 10/8, 15/8, 20/8}`): >>> 2.5 * (np.random.random_integers(5, size=(5,)) - 1) / 4. array([ 0.625, 1.25 , 0.625, 0.625, 2.5 ]) # random Roll two six sided dice 1000 times and sum the results: >>> d1 = np.random.random_integers(1, 6, 1000) >>> d2 = np.random.random_integers(1, 6, 1000) >>> dsums = d1 + d2 Display results as a histogram: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(dsums, 11, density=True) >>> plt.show() randn(d0, d1, ..., dn) Return a sample (or samples) from the "standard normal" distribution. .. note:: This is a convenience function for users porting code from Matlab, and wraps `standard_normal`. That function takes a tuple to specify the size of the output, which is consistent with other NumPy functions like `numpy.zeros` and `numpy.ones`. .. note:: New code should use the ``standard_normal`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. If positive int_like arguments are provided, `randn` generates an array of shape ``(d0, d1, ..., dn)``, filled with random floats sampled from a univariate "normal" (Gaussian) distribution of mean 0 and variance 1. A single float randomly sampled from the distribution is returned if no argument is provided. Parameters ---------- d0, d1, ..., dn : int, optional The dimensions of the returned array, must be non-negative. If no argument is given a single Python float is returned. Returns ------- Z : ndarray or float A ``(d0, d1, ..., dn)``-shaped array of floating-point samples from the standard normal distribution, or a single such float if no parameters were supplied. See Also -------- standard_normal : Similar, but takes a tuple as its argument. normal : Also accepts mu and sigma arguments. Generator.standard_normal: which should be used for new code. Notes ----- For random samples from :math:`N(\mu, \sigma^2)`, use: ``sigma * np.random.randn(...) + mu`` Examples -------- >>> np.random.randn() 2.1923875335537315 # random Two-by-four array of samples from N(3, 6.25): >>> 3 + 2.5 * np.random.randn(2, 4) array([[-4.49401501, 4.00950034, -1.81814867, 7.29718677], # random [ 0.39924804, 4.68456316, 4.99394529, 4.84057254]]) # random rand(d0, d1, ..., dn) Random values in a given shape. .. note:: This is a convenience function for users porting code from Matlab, and wraps `random_sample`. That function takes a tuple to specify the size of the output, which is consistent with other NumPy functions like `numpy.zeros` and `numpy.ones`. Create an array of the given shape and populate it with random samples from a uniform distribution over ``[0, 1)``. Parameters ---------- d0, d1, ..., dn : int, optional The dimensions of the returned array, must be non-negative. If no argument is given a single Python float is returned. Returns ------- out : ndarray, shape ``(d0, d1, ..., dn)`` Random values. See Also -------- random Examples -------- >>> np.random.rand(3,2) array([[ 0.14022471, 0.96360618], #random [ 0.37601032, 0.25528411], #random [ 0.49313049, 0.94909878]]) #random uniform(low=0.0, high=1.0, size=None) Draw samples from a uniform distribution. Samples are uniformly distributed over the half-open interval ``[low, high)`` (includes low, but excludes high). In other words, any value within the given interval is equally likely to be drawn by `uniform`. .. note:: New code should use the ``uniform`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- low : float or array_like of floats, optional Lower boundary of the output interval. All values generated will be greater than or equal to low. The default value is 0. high : float or array_like of floats Upper boundary of the output interval. All values generated will be less than or equal to high. The default value is 1.0. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``low`` and ``high`` are both scalars. Otherwise, ``np.broadcast(low, high).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized uniform distribution. See Also -------- randint : Discrete uniform distribution, yielding integers. random_integers : Discrete uniform distribution over the closed interval ``[low, high]``. random_sample : Floats uniformly distributed over ``[0, 1)``. random : Alias for `random_sample`. rand : Convenience function that accepts dimensions as input, e.g., ``rand(2,2)`` would generate a 2-by-2 array of floats, uniformly distributed over ``[0, 1)``. Generator.uniform: which should be used for new code. Notes ----- The probability density function of the uniform distribution is .. math:: p(x) = \frac{1}{b - a} anywhere within the interval ``[a, b)``, and zero elsewhere. When ``high`` == ``low``, values of ``low`` will be returned. If ``high`` < ``low``, the results are officially undefined and may eventually raise an error, i.e. do not rely on this function to behave when passed arguments satisfying that inequality condition. The ``high`` limit may be included in the returned array of floats due to floating-point rounding in the equation ``low + (high-low) * random_sample()``. For example: >>> x = np.float32(5*0.99999999) >>> x 5.0 Examples -------- Draw samples from the distribution: >>> s = np.random.uniform(-1,0,1000) All values are within the given interval: >>> np.all(s >= -1) True >>> np.all(s < 0) True Display the histogram of the samples, along with the probability density function: >>> import matplotlib.pyplot as plt >>> count, bins, ignored = plt.hist(s, 15, density=True) >>> plt.plot(bins, np.ones_like(bins), linewidth=2, color='r') >>> plt.show() choice(a, size=None, replace=True, p=None) Generates a random sample from a given 1-D array .. versionadded:: 1.7.0 .. note:: New code should use the ``choice`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- a : 1-D array-like or int If an ndarray, a random sample is generated from its elements. If an int, the random sample is generated as if it were ``np.arange(a)`` size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. replace : boolean, optional Whether the sample is with or without replacement. Default is True, meaning that a value of ``a`` can be selected multiple times. p : 1-D array-like, optional The probabilities associated with each entry in a. If not given, the sample assumes a uniform distribution over all entries in ``a``. Returns ------- samples : single item or ndarray The generated random samples Raises ------ ValueError If a is an int and less than zero, if a or p are not 1-dimensional, if a is an array-like of size 0, if p is not a vector of probabilities, if a and p have different lengths, or if replace=False and the sample size is greater than the population size See Also -------- randint, shuffle, permutation Generator.choice: which should be used in new code Notes ----- Setting user-specified probabilities through ``p`` uses a more general but less efficient sampler than the default. The general sampler produces a different sample than the optimized sampler even if each element of ``p`` is 1 / len(a). Sampling random rows from a 2-D array is not possible with this function, but is possible with `Generator.choice` through its ``axis`` keyword. Examples -------- Generate a uniform random sample from np.arange(5) of size 3: >>> np.random.choice(5, 3) array([0, 3, 4]) # random >>> #This is equivalent to np.random.randint(0,5,3) Generate a non-uniform random sample from np.arange(5) of size 3: >>> np.random.choice(5, 3, p=[0.1, 0, 0.3, 0.6, 0]) array([3, 3, 0]) # random Generate a uniform random sample from np.arange(5) of size 3 without replacement: >>> np.random.choice(5, 3, replace=False) array([3,1,0]) # random >>> #This is equivalent to np.random.permutation(np.arange(5))[:3] Generate a non-uniform random sample from np.arange(5) of size 3 without replacement: >>> np.random.choice(5, 3, replace=False, p=[0.1, 0, 0.3, 0.6, 0]) array([2, 3, 0]) # random Any of the above can be repeated with an arbitrary array-like instead of just integers. For instance: >>> aa_milne_arr = ['pooh', 'rabbit', 'piglet', 'Christopher'] >>> np.random.choice(aa_milne_arr, 5, p=[0.5, 0.1, 0.1, 0.3]) array(['pooh', 'pooh', 'pooh', 'Christopher', 'piglet'], # random dtype='>> np.random.bytes(10) b' eh\x85\x022SZ\xbf\xa4' #random randint(low, high=None, size=None, dtype=int) Return random integers from `low` (inclusive) to `high` (exclusive). Return random integers from the "discrete uniform" distribution of the specified dtype in the "half-open" interval [`low`, `high`). If `high` is None (the default), then results are from [0, `low`). .. note:: New code should use the ``integers`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- low : int or array-like of ints Lowest (signed) integers to be drawn from the distribution (unless ``high=None``, in which case this parameter is one above the *highest* such integer). high : int or array-like of ints, optional If provided, one above the largest (signed) integer to be drawn from the distribution (see above for behavior if ``high=None``). If array-like, must contain integer values size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. dtype : dtype, optional Desired dtype of the result. Byteorder must be native. The default value is int. .. versionadded:: 1.11.0 Returns ------- out : int or ndarray of ints `size`-shaped array of random integers from the appropriate distribution, or a single such random int if `size` not provided. See Also -------- random_integers : similar to `randint`, only for the closed interval [`low`, `high`], and 1 is the lowest value if `high` is omitted. Generator.integers: which should be used for new code. Examples -------- >>> np.random.randint(2, size=10) array([1, 0, 0, 0, 1, 1, 0, 0, 1, 0]) # random >>> np.random.randint(1, size=10) array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0]) Generate a 2 x 4 array of ints between 0 and 4, inclusive: >>> np.random.randint(5, size=(2, 4)) array([[4, 0, 2, 1], # random [3, 2, 2, 0]]) Generate a 1 x 3 array with 3 different upper bounds >>> np.random.randint(1, [3, 5, 10]) array([2, 2, 9]) # random Generate a 1 by 3 array with 3 different lower bounds >>> np.random.randint([1, 5, 7], 10) array([9, 8, 7]) # random Generate a 2 by 4 array using broadcasting with dtype of uint8 >>> np.random.randint([1, 3, 5, 7], [[10], [20]], dtype=np.uint8) array([[ 8, 6, 9, 7], # random [ 1, 16, 9, 12]], dtype=uint8) tomaxint(size=None) Return a sample of uniformly distributed random integers in the interval [0, ``np.iinfo(np.int_).max``]. The `np.int_` type translates to the C long integer type and its precision is platform dependent. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : ndarray Drawn samples, with shape `size`. See Also -------- randint : Uniform sampling over a given half-open interval of integers. random_integers : Uniform sampling over a given closed interval of integers. Examples -------- >>> rs = np.random.RandomState() # need a RandomState object >>> rs.tomaxint((2,2,2)) array([[[1170048599, 1600360186], # random [ 739731006, 1947757578]], [[1871712945, 752307660], [1601631370, 1479324245]]]) >>> rs.tomaxint((2,2,2)) < np.iinfo(np.int_).max array([[[ True, True], [ True, True]], [[ True, True], [ True, True]]]) standard_exponential(size=None) Draw samples from the standard exponential distribution. `standard_exponential` is identical to the exponential distribution with a scale parameter of 1. .. note:: New code should use the ``standard_exponential`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray Drawn samples. See Also -------- Generator.standard_exponential: which should be used for new code. Examples -------- Output a 3x8000 array: >>> n = np.random.standard_exponential((3, 8000)) exponential(scale=1.0, size=None) Draw samples from an exponential distribution. Its probability density function is .. math:: f(x; \frac{1}{\beta}) = \frac{1}{\beta} \exp(-\frac{x}{\beta}), for ``x > 0`` and 0 elsewhere. :math:`\beta` is the scale parameter, which is the inverse of the rate parameter :math:`\lambda = 1/\beta`. The rate parameter is an alternative, widely used parameterization of the exponential distribution [3]_. The exponential distribution is a continuous analogue of the geometric distribution. It describes many common situations, such as the size of raindrops measured over many rainstorms [1]_, or the time between page requests to Wikipedia [2]_. .. note:: New code should use the ``exponential`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- scale : float or array_like of floats The scale parameter, :math:`\beta = 1/\lambda`. Must be non-negative. size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``scale`` is a scalar. Otherwise, ``np.array(scale).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized exponential distribution. See Also -------- Generator.exponential: which should be used for new code. References ---------- .. [1] Peyton Z. Peebles Jr., "Probability, Random Variables and Random Signal Principles", 4th ed, 2001, p. 57. .. [2] Wikipedia, "Poisson process", https://en.wikipedia.org/wiki/Poisson_process .. [3] Wikipedia, "Exponential distribution", https://en.wikipedia.org/wiki/Exponential_distribution beta(a, b, size=None) Draw samples from a Beta distribution. The Beta distribution is a special case of the Dirichlet distribution, and is related to the Gamma distribution. It has the probability distribution function .. math:: f(x; a,b) = \frac{1}{B(\alpha, \beta)} x^{\alpha - 1} (1 - x)^{\beta - 1}, where the normalization, B, is the beta function, .. math:: B(\alpha, \beta) = \int_0^1 t^{\alpha - 1} (1 - t)^{\beta - 1} dt. It is often seen in Bayesian inference and order statistics. .. note:: New code should use the ``beta`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- a : float or array_like of floats Alpha, positive (>0). b : float or array_like of floats Beta, positive (>0). size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. If size is ``None`` (default), a single value is returned if ``a`` and ``b`` are both scalars. Otherwise, ``np.broadcast(a, b).size`` samples are drawn. Returns ------- out : ndarray or scalar Drawn samples from the parameterized beta distribution. See Also -------- Generator.beta: which should be used for new code. random(size=None) Return random floats in the half-open interval [0.0, 1.0). Alias for `random_sample` to ease forward-porting to the new random API. random_sample(size=None) Return random floats in the half-open interval [0.0, 1.0). Results are from the "continuous uniform" distribution over the stated interval. To sample :math:`Unif[a, b), b > a` multiply the output of `random_sample` by `(b-a)` and add `a`:: (b - a) * random_sample() + a .. note:: New code should use the ``random`` method of a ``default_rng()`` instance instead; please see the :ref:`random-quick-start`. Parameters ---------- size : int or tuple of ints, optional Output shape. If the given shape is, e.g., ``(m, n, k)``, then ``m * n * k`` samples are drawn. Default is None, in which case a single value is returned. Returns ------- out : float or ndarray of floats Array of random floats of shape `size` (unless ``size=None``, in which case a single float is returned). See Also -------- Generator.random: which should be used for new code. Examples -------- >>> np.random.random_sample() 0.47108547995356098 # random >>> type(np.random.random_sample()) >>> np.random.random_sample((5,)) array([ 0.30220482, 0.86820401, 0.1654503 , 0.11659149, 0.54323428]) # random Three-by-two array of random numbers from [-5, 0): >>> 5 * np.random.random_sample((3, 2)) - 5 array([[-3.99149989, -0.52338984], # random [-2.99091858, -0.79479508], [-1.23204345, -1.75224494]]) set_state(state) Set the internal state of the generator from a tuple. For use if one has reason to manually (re-)set the internal state of the bit generator used by the RandomState instance. By default, RandomState uses the "Mersenne Twister"[1]_ pseudo-random number generating algorithm. Parameters ---------- state : {tuple(str, ndarray of 624 uints, int, int, float), dict} The `state` tuple has the following items: 1. the string 'MT19937', specifying the Mersenne Twister algorithm. 2. a 1-D array of 624 unsigned integers ``keys``. 3. an integer ``pos``. 4. an integer ``has_gauss``. 5. a float ``cached_gaussian``. If state is a dictionary, it is directly set using the BitGenerators `state` property. Returns ------- out : None Returns 'None' on success. See Also -------- get_state Notes ----- `set_state` and `get_state` are not needed to work with any of the random distributions in NumPy. If the internal state is manually altered, the user should know exactly what he/she is doing. For backwards compatibility, the form (str, array of 624 uints, int) is also accepted although it is missing some information about the cached Gaussian value: ``state = ('MT19937', keys, pos)``. References ---------- .. [1] M. Matsumoto and T. Nishimura, "Mersenne Twister: A 623-dimensionally equidistributed uniform pseudorandom number generator," *ACM Trans. on Modeling and Computer Simulation*, Vol. 8, No. 1, pp. 3-30, Jan. 1998. get_state() Return a tuple representing the internal state of the generator. For more details, see `set_state`. Parameters ---------- legacy : bool, optional Flag indicating to return a legacy tuple state when the BitGenerator is MT19937, instead of a dict. Returns ------- out : {tuple(str, ndarray of 624 uints, int, int, float), dict} The returned tuple has the following items: 1. the string 'MT19937'. 2. a 1-D array of 624 unsigned integer keys. 3. an integer ``pos``. 4. an integer ``has_gauss``. 5. a float ``cached_gaussian``. If `legacy` is False, or the BitGenerator is not MT19937, then state is returned as a dictionary. See Also -------- set_state Notes ----- `set_state` and `get_state` are not needed to work with any of the random distributions in NumPy. If the internal state is manually altered, the user should know exactly what he/she is doing. seed(self, seed=None) Reseed a legacy MT19937 BitGenerator Notes ----- This is a convenience, legacy function. The best practice is to **not** reseed a BitGenerator, rather to recreate a new one. This method is here for legacy reasons. This example demonstrates best practice. >>> from numpy.random import MT19937 >>> from numpy.random import RandomState, SeedSequence >>> rs = RandomState(MT19937(SeedSequence(123456789))) # Later, you want to restart the stream >>> rs = RandomState(MT19937(SeedSequence(987654321))) This is an alias of `random_sample`. See `random_sample` for the complete documentation. This is an alias of `random_sample`. See `random_sample` for the complete documentation. a aQaSat(U{`P -UpK aD TU > T@W`= TZ7 T^. T`a* |U @% 7U a` VU V aF WI W0M` PV YV `V iV zV  oV V |V V{ V`j V] VM V? V 2 V`$ V0` V  VV@V @W@ [[b[pf[i [l[s[o[ЙnaLa(^pUa`m Qa,`Lx zA pz @z& 0z ȫ z H y> y p y@ y y$ y xP x h x P x! Px@ x" w w"` wp pw @w" w v' v  v! @v" v"Э u$ u,@ @u*0 u-P t% t pt @t$0 t Ю sЯ s s s `s(@ s% r!P r r  `r(0 r, q' q( `q# @q ` q# p  p!p p `p  @p p p p H n= nЧ `nG nL m ` m!ث m  m  m m m m mج `m  m0 l l& l5 |l xl` pl `l  SlP Hl  Bl@ ;l 6lH 0l (l0 l l  l l l8 l l k  k X ]` ]p \ \ \ Zm@ @Z( 0Z 0Z (Zx Z ` Y4 Y  Y H P ص Pر P C  B ` B BX B B B Bp `B%X B)( Bp Bh B B B B B A б A ȭ 51  5 58 5H 5 5 X x5  o5ئ i5` `5  P5 ȱ P5  K5 K5h `'  X' P'ȣ D'x >' >'x `  ZP P  P  J     0 `u X8 P P ` P  G 8 8 1     С  @  0   `   8   x  q h X aP a ]( V T P P H HȮ      x p ` Sh H  @@ @أ ;P 0      ? ش       x   @ P 5 1 (  8  h #x    h   p ` lH l ` p ` ح @  h  0 y y p ة g g bx P` PH G      Њ Ŋx Ŋ` X 8 ~ H ~ P ~ X q3 0 eqH eq be@ aЩ a ax a8 a'0 a"` ha  Pa Ha` ` p`  j` j` P` F`@ F` P XP 8 XP  Kh K K0 K8 A X A Aس ~A( ~Aخ 3 Ȭ 3#Ь P3 03ؠ 3 3 3p 3 2؟ 2 2 2د .q8 h. h. #E  # #ȯ u  p   `  H HH p     yH p  p   H ` X0 Pȟ @ Ы 0  ' ' "  p h    ذ X }x ` а `  ?@  8 Ȱ   `0 Z( ZȢ U P Kȷ @  0 0(    8 в    ا        QH  ( p @!У 8( 0  и   h     p ( S    h [  (     x   0 П ( `      H  x : 0 q q8 h 0 `x ` ~ @ `~/p @~ 4~ /~p /~( s 8jPKmqDXj  @T S c a^@HR F^0E`S a Z Ȣ 0 Ȣ 6a0104e89977f4c46abc3ba32d22c5d5a842ac.debugN.shstrtab.note.gnu.property.note.gnu.build-id.gnu.hash.dynsym.dynstr.gnu.version.gnu.version_r.rela.dyn.rela.plt.init.plt.got.plt.sec.text.fini.rodata.eh_frame_hdr.eh_frame.init_array.fini_array.data.rel.ro.dynamic.got.plt.data.bss.gnu_debuglink  $1o ; H0C??hKoPZPZXoX^X^g^^mqBP{v  QKK PP {    00PG<,<,<,0 =->.@0F6Y    4ԏ