>hQXSSKrSSKrSSKrSSKJr SSKJs J r SSK J r "SS\R5r"SS\5r"SS \5r"S S \5r"S S \R$5r"SS\ R(5r\ R,"\\5 Sr"SS\5r\r\r\r\rg)N)model)ConvergenceWarningc2^\rSrSrSrU4SjrSrSrU=r$)_DimReductionRegression z: A base class for dimension reduction regression methods. c (>[TU]"X40UD6 gN)super__init__selfendogexogkwargs __class__s pC:\Users\julio\OneDrive\Documentos\Trabajo\Ideas Frescas\venv\Lib\site-packages\statsmodels/regression/dimred.pyr _DimReductionRegression.__init__s //c[R"UR5nURUSS24nX3R S5-n[R "UR U5URS- n[RRU5n[RRXSR 5R nX0l XPl [R"X15UlgNr)npargsortrrmeandotTshapelinalgcholeskysolvewexog_covxr array_split _split_wexog)r n_sliceiixcovxcovxrs r_prep_DimReductionRegression._prepsZZ # IIb!e  VVAYvvacc1~ * ""4( IIOOE33 ' ) )  NN16r)r!r#r ) __name__ __module__ __qualname____firstlineno____doc__r r)__static_attributes__ __classcell__rs@rrr s077rrc<\rSrSrSrS SjrSrSrS SjrSr g) SlicedInverseReg%a Sliced Inverse Regression (SIR) Parameters ---------- endog : array_like (1d) The dependent variable exog : array_like (2d) The covariates References ---------- KC Li (1991). Sliced inverse regression for dimension reduction. JASA 86, 316-342. c N[U5S:aSn[R"U5 URRSU-nUR U5 UR Vs/sHoURS5PM nnUR Vs/sHoURSPM nn[R"U5n[R"U5n[R"URUSS2S4U-5UR5- n[RRU5up[R"U *5n Xn U SS2U 4n [RR!UR"RU 5n [%X U S9n ['U 5$s snfs snf)z Estimate the EDR space using Sliced Inverse Regression. Parameters ---------- slice_n : int, optional Target number of observations per slice rz1SIR.fit does not take any extra keyword argumentsNeigs)lenwarningswarnrrr)r#rrasarrayrrsumreighrrr!DimReductionResultsDimReductionResultsWrapper)r slice_nrmsgr$zmnnmncabjjparamsresultss rfitSlicedInverseReg.fit6sB v;?EC MM# ))//!$/ 7!%!2!2 3!2AffQi!2 3!%!2!2 3!2AWWQZ!2 3 ZZ^ JJqMffRTT1QW:?+aeeg5yy~~c" ZZ^ E aeH2%d;)'224 3s #F F"cURnURnURnURnSn[R "XUR 45n[UR 5HFn[R"URUSS2U45nU[R"X-5- nMH [R"X15n [RRU 5up[R"U [R"U RUR55n URU - n U[R"X]U -RS55- nU$r)k_vars_covx _slice_means _slice_propsrreshapendimrangerpen_matr=rqrr)r Apr'rDphvkucovxaq_qdqus r_regularized_objective'SlicedInverseReg._regularized_objective[s KKzz        JJqdii. )tyy!At||Qq!tW-A  A" tyy||E" VVArvvacc244( ) TTBY RVVBb a( ))rc URnURnURnURnURnUR nUR X#45nS[R"URR[R"URU55-nUR X#45n[R"XA5n [R"XI5n [R"U RU 5n [RRU 5n [R"X#45n [RRXR5nS/X#--n[U5GHGn[U5GH3nU S-n SU UU4'[R"U RU 5nUUR- n[R"U [R"UU 55*n[R"[R"XM5U5nU[R"U [R"UU R55- nU[R"U [RRU [R"U RU555- nUUUU-U-'GM6 GMJ [RRU [R"U RUR55nU[R"U U5R- n[U5HnUUSS24nUUSS24n[U5Hbn[U5HPn[R"U[R"UUU-U-U55nUUU4==SUU-U--ss'MR Md M UR!5$)Nr)rOrTrPr$rQrRrSrrrVrrinvzerosrrUravel)r rXrYrTr'r$rDrZgrr^covx2aQQijmqcvftr_rumatfmatchcuir]r[fs r_regularized_grad"SlicedInverseReg._regularized_gradss KKyyzz,,       IIqi  t||Q(?@ @ IIqi t$ FF577E " YY]]1  XXqi iiooa)Vqx qA4[a1a4vvfhh+r266$#344vvbffT.4ubffT577&;<<ubiiooad9K&LMM!%1T6A:!YY__Quww 5 6 "&&#%% %wA1a4A1a4A1XtAq"&&AdFQJ"; >zz.$7 **Y+))//!$/ZZ # IIb!e  VVAYvvacc{^^A/ !+ ,AffQi ,!+ ,AWWQZ , ZZ^ JJqMK jjm      XXt{{D12F%'VVD\F1T61T6> "F!!!$,I o%!F%fd.I.I&*&<&)'22A- ,s 6J7J<)rPrQrRrOr$rTrV)r})rgNr}dgMbP?) r+r,r-r.r/rLrcryrr0rrr4r4%s( #3J0-^IL!c3rr4c\rSrSrSrSrSrg)PrincipalHessianDirectionsia Principal Hessian Directions (PHD) Parameters ---------- endog : array_like (1d) The dependent variable exog : array_like (2d) The covariates Returns ------- A model instance. Call `fit` to obtain a results instance, from which the estimated parameters can be obtained. References ---------- KC Li (1992). On Principal Hessian Directions for Data Visualization and Dimension Reduction: Another application of Stein's lemma. JASA 87:420. c URSS5nURURR5- nURURRS5- nU(a(SSKJn U"X45R 5nURn[R"SX4U5nU[U5-n[R"UR5n[RRX5n [RRU 5up[R "[R""U 5*5n Xn U SS2U 4n [%X U S9n['U5$)a? Estimate the EDR space using PHD. Parameters ---------- resid : bool, optional If True, use least squares regression to remove the linear relationship between each covariate and the response, before conducting PHD. Returns ------- A results instance which can be used to access the estimated parameters. residFr)OLSz i,ij,ik->jkNr7)rrrr#statsmodels.regression.linear_modelrrLrreinsumr9rrrreigrabsr?r@)r rryr&rrrcmcxcbrGrHrIrJrKs rrLPrincipalHessianDirections.fits" 7E* JJ* * II q) )  ?A AA YY}aA . c!f  VVACC[ YY__R $yy}}R  ZZ # E1b5%d;)'22rrN)r+r,r-r.r/rLr0rrrrrs ,'3rrc2^\rSrSrSrU4SjrSrSrU=r$)SlicedAverageVarianceEstimationiIa Sliced Average Variance Estimation (SAVE) Parameters ---------- endog : array_like (1d) The dependent variable exog : array_like (2d) The covariates bc : bool, optional If True, use the bias-corrected CSAVE method of Li and Zhu. References ---------- RD Cook. SAVE: A method for dimension reduction and graphics in regression. http://www.stat.umn.edu/RegGraph/RecentDev/save.pdf Y Li, L-X Zhu (2007). Asymptotics for sliced average variance estimation. The Annals of Statistics. https://arxiv.org/pdf/0708.0462.pdf c l>[[U] "X40UD6 SUlSU;aUSSLaSUlggg)NFbcT)r SAVEr rr s rr (SlicedAverageVarianceEstimation.__init__as> dD"59&9 6>fTld2DG3>rc URSS5nURRSU-nURU5 URVs/sH#n[ R "UR5PM% nnURVs/sHoDRSPM nnURRSnUR(dZSn[Xe5H9up[ R"U5U - n X[ R"X5-- nM; U[U5-nGOGSn UHn U [ R"X5- n M U [U5-n SnURHonXRS5- n[URS5H=nUUSS24n[ R "UU5nU[ R"UU5- nM? Mq XRRS-n[ R"U5n XS- -U S- S-S-- nU S- U S- S-S-- nUU -UU-- n[ R"U5S[#U5-[U5- - U-n[ R$R'U5unn[ R("U*5nUUnUSS2U4n[ R$R+UR,RU5n[/UUUS9n5$s snfs snf)zc Estimate the EDR space. Parameters ---------- slice_n : int Number of observations per slice rA2rrgNrfr7)rrrr)r#rrrr rziprrr9rrUouterr=rr>rrr!r?r@)r rrAr$rCcvnsrYvmwcvxicvavcvnr&rrrwr]mk1k2av2rGrHrIrJrKs rrL#SlicedAverageVarianceEstimation.fiths**Y+))//!$/ 7#'#4#4 5#4abffQSSk#4 5"&"3"3 4"3Qggaj"3 4 JJ  Q wwBb+ffQi#o"&&***& #b'MB BbffQl" #b'MBB&&q Mqwwqz*A!Q$AAA"&&A,&B+' ))//!$ $B A!eQ Q/Ba%QUQJN+Br'BG#CQR[3r722S8Byy~~b!1 ZZ^ bE aeH2%dF;)'22[6 4s *K>L)r) r+r,r-r.r/r rLr0r1r2s@rrrIs.?3?3rrc,^\rSrSrSrU4SjrSrU=r$)r?ia2 Results class for a dimension reduction regression. Notes ----- The `params` attribute is a matrix whose columns span the effective dimension reduction (EDR) space. Some methods produce a corresponding set of eigenvalues (`eigs`) that indicate how much information is contained in each basis direction. c0>[TU]X5 X0lgr )r r r8)r rrJr8rs rr DimReductionResults.__init__s   rr7)r+r,r-r.r/r r0r1r2s@rr?r?s rr?c \rSrSrSS0r\rSrg)r@irJcolumnsrN)r+r,r-r._attrs _wrap_attrsr0rrrr@r@s)FKrr@c^^^^URupVUR5nU"U5nSn[U5GHn U"U5n U [R"X5U-[R"X5- -n [R "[R "X-55U:aSn OU RXV45n [RRU S5ummmURXV45n [R"U TR5mUUUU4Sjn SnUS:dMU "U*5nU"U5nUU:aUnUnGM US-nUS:aM+GM URXV45nXU4$)a Minimize a function on a Grassmann manifold. Parameters ---------- params : array_like Starting value for the optimization. fun : function The function to be minimized. grad : function The gradient of fun. maxiter : int The maximum number of iterations. gtol : float Convergence occurs when the gradient norm falls below this value. Returns ------- params : array_like The minimizing value for the objective function. fval : float The smallest achieved value of the objective function. cnvrg : bool True if the algorithm converged to a limit point. Notes ----- `params` is 2-d, but `fun` and `grad` should take 1-d arrays `params.ravel()` as arguments. Reference --------- A Edelman, TA Arias, ST Smith (1998). The geometry of algorithms with orthogonality constraints. SIAM J Matrix Anal Appl. http://math.mit.edu/~edelman/publications/geometry_of_algorithms.pdf FTrc>T[R"TU-5-T[R"TU-5--n[R"UT5R 5$r )rcossinrrj)tpapa0sr]vts rgeo_grass_opt..geosJrvva!e}$q266!a%='88B66"b>'') )rg@g|=rf) rrjrUrrrr=rSrsvdr)rJfungradrrrYdf0rr`rgmparamsmrsteprf1rrr]rs @@@@rrrsLL <[T U]X5 //pT[R"URUR S9nUR UR5HLupxURUR5R5 URURS5 MN [U5Ul Sn [U5HupXU XZ-- n M XR-n XlX@lXPlX0lg)N)indexr)r r pd DataFramerrgroupbyrappendrvaluesrr9nobs enumeratecovmcovsrdim) r rrrrrdfr`r[rrwrs rr CovarianceReduction.__init__@s %rb \\$))4:: 6JJrxx(DA KK ' IIaggaj !)J dODA GbeO #D$   rclURRSnURX R45n[R "UR [R "URU55n[RRU5upVURU-S- n[UR5Hrup[R "UR [R "XC55n[RRU5upVXpRUU-S- -nMt U$)z Evaluate the log-likelihood Parameters ---------- params : array_like The projection matrix used to reduce the covariances, flattened to 1d. Returns the log-likelihood. rrf) rrrSrrrrrslogdetrrrr) r rJrYprojrr`ldetrxjs rloglikeCovarianceReduction.loglikeWs IIOOA ~~q((m, FF466266$))T2 3))##A& II q dii(DAtvvrvva/Aii''*GA d"Q& &A) rc*URRSnURX R45n[R "UR [R "URU55n[R "URU5nUR[RRXER 5R -n[UR5Hupx[R "UR [R "X55n[R "X5nX`RU[RRXER 5R --nM UR5$)z Evaluate the score function. Parameters ---------- params : array_like The projection matrix used to reduce the covariances, flattened to 1d. Returns the score function evaluated at 'params'. r)rrrSrrrrrrrrrrrj) r rJrYrc0cPrrrs rscoreCovarianceReduction.scorers IIOOA ~~q((m, VVDFFBFF499d3 4 VVDIIt $ II DD133 3dii(DAq0BB biioob$$7999 9A) wwyrc$^TRRSnTRnUc9[R"XE45n[R "U5USU2SU24'UnOUn[ UU4SjU4SjUU5upgnUS-nU(dkTRUR55n [R"[R"X-55n SU -n [R"U [5 [TUSS9n X|l[!U 5$)a~ Fit the covariance reduction model. Parameters ---------- start_params : array_like Starting value for the projection matrix. May be rectangular, or flattened. maxiter : int The maximum number of gradient steps to take. gtol : float Convergence criterion for the gradient norm. Returns ------- A results instance that can be used to access the fitted parameters. rNc(>TRU5*$r )rr&r s r)CovarianceReduction.fit..s4<<?:Jrc(>TRU5*$r )rrs rrrs4::a=.rz/CovReduce optimization did not converge, |g|=%fr7)rrrrrirrrrjrr=r:r;rr?llfr@) r r|rrrYrrJrrrrrBrKs ` rrLCovarianceReduction.fits( IIOOA  HH  XXqf%F!vvayF1Q3!8 F!F(0J(@'(,.U r  6<<>*A'BCbHC MM#1 2%dF> )'22r)rrrrr)Ng-C6?) r+r,r-r.r/r rrrLr0r1r2s@rrrs"%N.66-3-3rr)r:numpyrpandasrstatsmodels.baserstatsmodels.base.wrapperbasewrapperwrapstatsmodels.tools.sm_exceptionsrModelrr4rrResultsr?ResultsWrapperr@populate_wrapperrrSIRPHDrCORErrrrs"''>7ekk74`3.`3F>3!8>3B^3&=^3B%--&!4!4 0)+Nbb31b3L &r