M/Ph(ObddlZddlZddlZddlmZddlmcm Z ddl m Z Gddej ZGddeZGdd eZGd d eZGd d ejZGdde jZe jeedZGddeZeZeZeZeZdS)N)model)ConvergenceWarningc(eZdZdZfdZdZxZS)_DimReductionRegressionzB A base class for dimension reduction regression methods. c >tj||fi|dSN)super__init__selfendogexogkwargs __class__s ]/var/www/html/test/jupyter/venv/lib/python3.11/site-packages/statsmodels/regression/dimred.pyr z _DimReductionRegression.__init__s*///////ctj|j}|j|ddf}||dz}tj|j||jdz }tj |}tj ||jj}||_ ||_ tj |||_dSNr)npargsortr rmeandotTshapelinalgcholeskysolvewexog_covxr array_split _split_wexog)r n_sliceiixcovxcovxrs r_prepz_DimReductionRegression._prepsZ # # Ib!!!e  QVVAYYvac1~~ * ""4(( IOOE13 ' ' )  N1g66r)__name__ __module__ __qualname____doc__r r' __classcell__rs@rrr sQ000007777777rrc2eZdZdZd dZdZdZ d d ZdS) SlicedInverseRega0 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 t|dkrd}tj||jjd|z}||d|jD}d|jD}tj|}tj|}tj |j |dddf|z| z }tj |\}} tj| } || }| dd| f} tj |jj | } t#|| |} t%| S)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 argumentsc8g|]}|dSrr.0zs r z(SlicedInverseReg.fit..Js" 3 3 3AaffQii 3 3 3rc(g|]}|jdSr3rr5s rr8z(SlicedInverseReg.fit..Ks 3 3 3AQWQZ 3 3 3rNeigs)lenwarningswarnrrr'r!rasarrayrrsumreighrrrDimReductionResultsDimReductionResultsWrapper) r slice_nrmsgr"mnnmncabjjparamsresultss rfitzSlicedInverseReg.fit6sB v;;??EC M#   )/!$/ 7 3 3!2 3 3 3 3 3!2 3 3 3 Z^^ JqMMfRT1QQQW:?++aeegg5y~~c""1 Z^^ bE aaaeH22%dF;;;)'222rcf|j}|j}|j}|j}d}t j|||jf}t|jD]@}t j|j |dd|f}|t j ||zz }At j||} tj | \} } t j| t j| j |j } |j | z } |t j|| | z dz }|Sr)k_vars_covx _slice_means _slice_propsrreshapendimrangerpen_matrArqrr)r Apr%rGphvkucovxaq_qdqus r_regularized_objectivez'SlicedInverseReg._regularized_objective[s Kz      Jq1di. ) )ty!!  At|Qqqq!tW--A A AAtQy||E""1 VArvac24(( ) ) TBY RVBb a(( ) ))rc |j}|j}|j}|j}|j}|j}|||f}dtj|j j tj|j |z}|||f}tj||} tj|| } tj| j | } tj | } tj ||f} tj | | j }dg||zz}t|D]!}t|D] }| dz} d| ||f<tj| j | }||j z }tj| tj||  }tjtj|| |}|tj| tj|| j z }|tj| tj | tj| j |z }||||z|z<#tj | tj| j |j }|tj| |j z }t|D]}||ddf}||ddf}t|D]f}t|D]T}tj|tj|||z|z|}|||fxxd||z|zzcc<Ug|S)Nr)rQrVrRr"rSrTrUrrrXrrinvzerosrrWravel)r rZr[rVr%r"rGr\grr`covx2aQQijmqcvftrarumatfmatchcuir_r]fs r_regularized_gradz"SlicedInverseReg._regularized_gradss Kyz,     IIq$i  t|Q(?(?@@ @ IIq$i tQe$$ F57E " " Y]]1   Xq$i iooa))Vq4x q & &A4[[ & &a1a4vfh++r26$#3#3444vbfT2..44ubfT57&;&;<<<ubiooad9K9K&L&LMMM!%1T6A: &Y__Quw 5 5 6 6 "&##% %w . .A1aaa4A1aaa4A1XX . .t..Aq"&AdFQJ";";<.s" , , ,AaffQii , , ,rc(g|]}|jdSr3r:r5s rr8z4SlicedInverseReg.fit_regularized..s , , ,AQWQZ , , ,rrhz1Shape of start_params is not compatible with ndimz,SIR.fit_regularized did not converge, |g|=%fr;)!r=r>r? ValueErrorgetrrrrr rcovrr r@rArTrVrQrXrRr"rSrjeye _grass_optrerzrksqrtrrCrD)r rVrXrEmaxitergtolrrFr~r"r#r$r% split_exogrGrHrMrbcnvrgggnrNs rfit_regularizedz SlicedInverseReg.fit_regularizedshT v;;??GC M#    ?=>> >zz.$77 **Y++)/!$/Z # # Ib!!!e  QVVAYYvac{{^Aw// , , , , , , , , , , Z^^ JqMMK jm      Xt{D122F%'VD\\F1T61T6> "FF!!$,,I oo%!F%fd.I&*&>>)'222r)r0)rhNr0r{r|)r(r)r*r+rOrerzrrrr/r/%sv #3#3#3#3J0---^IL!c3c3c3c3c3c3rr/ceZdZdZdZdS)PrincipalHessianDirectionsa Principal Hessian Directions (PHD) Parameters ---------- endog : array_like (1d) The dependent variable exog : array_like (2d) The covariates Returns ------- A model instance. Call `fit` to obtain a results instance, from which the estimated parameters can be obtained. References ---------- KC Li (1992). On Principal Hessian Directions for Data Visualization and Dimension Reduction: Another application of Stein's lemma. JASA 87:420. c |dd}|j|jz }|j|jdz }|r+ddlm}|||}|j}tj d|||}|t|z}tj |j }tj ||} tj | \} } tjtj|  } | | } | dd| f} t%|| | }t'|S)a Estimate the EDR space using PHD. Parameters ---------- resid : bool, optional If True, use least squares regression to remove the linear relationship between each covariate and the response, before conducting PHD. Returns ------- A results instance which can be used to access the estimated parameters. residFr)OLSz i,ij,ik->jkNr;)rr rr#statsmodels.regression.linear_modelrrOrreinsumr=rrrreigrabsrCrD)r rryr$rrscmcxcbrJrKrLrMrNs rrOzPrincipalHessianDirections.fits5" 7E** J** * I q)) )   ? ? ? ? ? ?Aq AA Y}aA . . c!ff  VAC[[ Y__R $ $y}}R  1 Z # # bE111b5%dF;;;)'222rN)r(r)r*r+rOrrrrrs-,'3'3'3'3'3rrc(eZdZdZfdZdZxZS)SlicedAverageVarianceEstimationa] 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 tt|j||fi|d|_d|vr|ddur d|_dSdSdS)NFbcT)r SAVEr rr s rr z(SlicedAverageVarianceEstimation.__init__as`"dD"5$99&999 6>>fTld22DGGG >22rc \|dd}|jjd|z}||d|jD}d|jD}|jjd}|js^d}t||D]7\}} tj || z } ||tj | | zz }8|t|z}n>d} |D]} | tj | | z } | t|z} d} |jD]p}|| dz }t|jdD];}||ddf}tj||}| tj ||z } .zs" 5 5 5abfQSkk 5 5 5rc(g|]}|jdSr3r:r5s rr8z7SlicedAverageVarianceEstimation.fit..{s 4 4 4Qagaj 4 4 4rrhNrgr;)rrrr'r!rrziprrrr=rrWouterrArrBrrrrrCrD)r rrEr"cvnsr[vmwcvxicvavcvnr$rsrxr_mk1k2av2rJrKrLrMrNs rrOz#SlicedAverageVarianceEstimation.fiths**Y++)/!$/ 7 5 54#4 5 5 5 4 4$"3 4 4 4 J Q w 9Bb"++ + +3fQii#oa"&c**** #b''MBB B # #bfQll" #b''MBB& ' 'q Mqwqz**''A!QQQ$AAA"&A,,&BB' $)/!$ $B Aa!eQ Q/Ba%QUQJN+Br'BG#CQR[3r7722S8By~~b!!1 Z^^ bE aaaeH22%dF;;;)'222r)r(r)r*r+r rOr,r-s@rrrIsQ.?3?3?3?3?3?3?3rrc"eZdZdZfdZxZS)rCaV Results class for a dimension reduction regression. Notes ----- The `params` attribute is a matrix whose columns span the effective dimension reduction (EDR) space. Some methods produce a corresponding set of eigenvalues (`eigs`) that indicate how much information is contained in each basis direction. cZt||||_dSr)r r r<)r rrMr<rs rr zDimReductionResults.__init__s/ V    r)r(r)r*r+r r,r-s@rrCrCsB  rrCceZdZddiZeZdS)rDrMcolumnsN)r(r)r*_attrs _wrap_attrsrrrrDrDs!)FKKKrrDc|j\}}|}||}d}t|D]} ||} | tj| ||ztj||z z} tjtj| | z|krd}n| ||f} tj | d\|||f} tj| j fd} d}|dkr-| | }||}||kr|}|}n |dz}|dk-|||f}|||fS)a Minimize a function on a Grassmann manifold. Parameters ---------- params : array_like Starting value for the optimization. fun : function The function to be minimized. grad : function The gradient of fun. maxiter : int The maximum number of iterations. gtol : float Convergence occurs when the gradient norm falls below this value. Returns ------- params : array_like The minimizing value for the objective function. fval : float The smallest achieved value of the objective function. cnvrg : bool True if the algorithm converged to a limit point. Notes ----- `params` is 2-d, but `fun` and `grad` should take 1-d arrays `params.ravel()` as arguments. Reference --------- A Edelman, TA Arias, ST Smith (1998). The geometry of algorithms with orthogonality constraints. SIAM J Matrix Anal Appl. http://math.mit.edu/~edelman/publications/geometry_of_algorithms.pdf FTrctj|zztj|zzz}tj|Sr)rcossinrrk)tpapa0sr_vts rgeoz_grass_opt..geosMrva!e}}$q26!a%=='88B6"b>>'')) )rg@g|=rg) rrkrWrrrrArUrsvdr)rMfungradrrr[df0rrbrgmparamsmrsteprf1rrr_rs @@@@rrrsL z)CovarianceReduction.fit..s4<<??:Jrc0| Sr)rrs rrz)CovarianceReduction.fit..s4::a==.rz/CovReduce optimization did not converge, |g|=%fr;)rrrrrjrrrrkrrAr>r?rrCllfrD) r r~rrr[rrMrrrrrFrNs ` rrOzCovarianceReduction.fits#( IOA  H  Xq!f%%F!vayyF1Q3!8 FF!F(0J0J0J0J(@(@(@(@'(,..U r  3 6<<>>**AA''BCbHC M#1 2 2 2%dF>>> )'222r)Nrr) r(r)r*r+r rrrOr,r-s@rrrsu%%N.66-3-3-3-3-3-3-3-3rr)r>numpyrpandasrstatsmodels.baserstatsmodels.base.wrapperbasewrapperwrapstatsmodels.tools.sm_exceptionsrModelrr/rrResultsrCResultsWrapperrDpopulate_wrapperrrSIRPHDrCORErrrrs""""""'''''''''>>>>>>77777ek7774`3`3`3`3`3.`3`3`3F>3>3>3>3>3!8>3>3>3B^3^3^3^3^3&=^3^3^3B%-&!4 0)+++NNNbb3b3b3b3b31b3b3b3L &r