M/PhyGdZddlZddlZddlmZddlmZ ddlZdZ n #e $rdZ YnwxYwddl m Z e e je je je je je je je je je je je je j Zd Zd ZGd d ZGddZGddZ dZ!dZ" ddZ#dS)zh Module containing the base object for multivariate kernel density and regression, plus some utilities. N)optimize) mquantilesTF)kernels) wangryzinaitchisonaitkengaussianaitchison_aitken_reg wangryzin_reggauss_convolutionwangryzin_convolutionaitchisonaitken_convolution gaussian_cdfaitchisonaitken_cdf wangryzin_cdf d_gaussiantricubectj|d}t|ddjd}t|ddjd}||z dz }tj||}|S)z1Compute minimum of std and IQR for each variable.raxisg?g?g/$?)npstdrdataminimum)rs1q75q25s2 dispersions f/var/www/html/test/jupyter/venv/lib/python3.11/site-packages/statsmodels/nonparametric/_kernel_base.py_compute_min_std_IQRr!#ss 1   B T4a ( ( ( -a 0C T4a ( ( ( -a 0C )u BB##J c H| r.tj||d|ddf} n|| d| dddf} |dkr,ddlm} | d}| | ||t d}n|d krHdd lm}| \}}}| ddd|f}| dd|df}||||||t d}nw|d krbdd lm}| \}}}t| dddfd}t| ddddf|}||||||t d }ntd|d kr| ddddf} t| }||d||zz zz}|d||zz z||<|d||zz z||<|j |z }|j }||fS)z"Compute bw on subset of data. Called from ``GenericKDE._compute_efficient_*``. Notes ----- Needs to be outside the class in order for joblib to be able to pickle it. NrrKDEMultivariate)r$F) efficient)bwdefaultsKDEMultivariateConditional)r( KernelReg)r))endogexogreg_typevar_typer&r'zdclass_type not recognized, should be one of {KDEMultivariate, KDEMultivariateConditional, KernelReg}) rrandomshufflekernel_densityr$EstimatorSettingsr(kernel_regressionr) _adjust_shape ValueErrorr!r&) class_typerr&codon_cvarsix_ordix_unordn_sub class_vars randomizeboundsub_datar$r- sub_modelr(k_depdep_type indep_typer*r+r)k_varsr,rfctsample_scale_subbw_subs r _compute_subsetrJ-s. $ ?aq)111,-&&&333333a=#OHh2!2U!C!C!CEEE 3 3 3>>>>>>&0#xFUF#566 "..udH 2(9E(J(J(JLLL { " "000000%/"&(hqqq!tna00Xaaae_f55IEx'/B'85'I'I'IKKK LMM M[  AAAqrrE?%h//J usgl34 4CC7R<01CM#2./CK |c) \F V ##r"cNeZdZdZdZdZdZdZdZdZ dZ d Z d Z d Z d S) GenericKDEzG Base class for density estimation and regression KDE classes. c|d}t|tsd|_tj|}n4||_|dkr|j}n|dkr|j}n|j}|}|S)a Computes the bandwidth of the data. Parameters ---------- bw : {array_like, str} If array_like: user-specified bandwidth. If a string, should be one of: - cv_ml: cross validation maximum likelihood - normal_reference: normal reference rule of thumb - cv_ls: cross validation least squares Notes ----- The default values for bw is 'normal_reference'. Nnormal_referenceuser-specifiedcv_ml) isinstancestr _bw_methodrasarray_normal_reference_cv_ml_cv_ls)selfr&resbwfuncs r _compute_bwzGenericKDE._compute_bwgs~$ :#B"c"" .DO*R..CC!DO'''/w&((C r"c t|S)a Computes the measure of dispersion. The minimum of the standard deviation and interquartile range / 1.349 Notes ----- Reimplemented in `KernelReg`, because the first column of `data` has to be removed. References ---------- See the user guide for the np package in R. In the notes on bwscaling option in npreg, npudens, npcdens there is a discussion on the measure of dispersion )r!)rXrs r _compute_dispersionzGenericKDE._compute_dispersions"$D)))r"cdS)zpHelper method to be able to pass needed vars to _compute_subset. Needs to be implemented by subclasses.NrXs r _get_class_vars_typezGenericKDE._get_class_vars_types  r"c d_ttr_n d_Sj}jt jjj dddtj\}j r dgj z nDfdt|zD |zdkr ||zz |fj rj nt }t!j|jf}t!j|jf}\ t(rHt+jj  f d t|D}nLg}t|D]:}|t1 j | ;t|D],}||d||ddf<||d ||ddf<-} jr t jn t j} | |d } | | z|d zz zz| |d zz zz<| |d zz zz<jrt!j|d S)a\ Computes the bandwidth by estimating the scaling factor (c) in n_res resamples of size ``n_sub`` (in `randomize` case), or by dividing ``nobs`` into as many ``n_sub`` blocks as needed (if `randomize` is False). References ---------- See p.9 in socserv.mcmaster.ca/racine/np_faq.pdf NrNrOcc*g|]}|z|dzzfS)rr_).0ir=s r z1GenericKDE._compute_efficient..s*OOOQq5y1Q3%-0OOOr"r)n_jobsc3 K|]<}tjt  j| V=dSN)joblibdelayedrJr?)rfrgboundsr&r7r>r8rr9r;r<r:r=rXs r z0GenericKDE._compute_efficient..sq6*6*0//b"b'68:t~vayBB6*6*6*6*6*6*r"rrr.r/)rSrQrRnobsr=copydeepcopyr data_typecount _get_type_posr?n_resrangeappendlenremptyrFra has_joblibrlParallelrirJr] return_medianmedianmeanreturn_only_bw)rXr&rp_n_blocks sample_scaleonly_bwrYrgs order_funcm_scalernr7r>r8rr9r;r<r:r=s`` @@@@@@@@@@r _compute_efficientzGenericKDE._compute_efficients] :0DO b#    DOO.DOIy }TY''.&&s++  +DN;;68 > ;Vdj(FFOOOO% :N:NOOOFe|a tdUl2D9:::!%@4::S[[x4; 788 (Hdk233!%!:!:!0"""""r"rLceZdZdZ d dZdS) r3a Object to specify settings for density estimation or regression. `EstimatorSettings` has several properties related to how bandwidth estimation for the `KDEMultivariate`, `KDEMultivariateConditional`, `KernelReg` and `CensoredKernelReg` classes behaves. Parameters ---------- efficient : bool, optional If True, the bandwidth estimation is to be performed efficiently -- by taking smaller sub-samples and estimating the scaling factor of each subsample. This is useful for large samples (nobs >> 300) and/or multiple variables (k_vars > 3). If False (default), all data is used at the same time. randomize : bool, optional If True, the bandwidth estimation is to be performed by taking `n_res` random resamples (with replacement) of size `n_sub` from the full sample. If set to False (default), the estimation is performed by slicing the full sample in sub-samples of size `n_sub` so that all samples are used once. n_sub : int, optional Size of the sub-samples. Default is 50. n_res : int, optional The number of random re-samples used to estimate the bandwidth. Only has an effect if ``randomize == True``. Default value is 25. return_median : bool, optional If True (default), the estimator uses the median of all scaling factors for each sub-sample to estimate the bandwidth of the full sample. If False, the estimator uses the mean. return_only_bw : bool, optional If True, the estimator is to use the bandwidth and not the scaling factor. This is *not* theoretically justified. Should be used only for experimenting. n_jobs : int, optional The number of jobs to use for parallel estimation with ``joblib.Parallel``. Default is -1, meaning ``n_cores - 1``, with ``n_cores`` the number of available CPU cores. See the `joblib documentation `_ for more details. Examples -------- >>> settings = EstimatorSettings(randomize=True, n_jobs=3) >>> k_dens = KDEMultivariate(data, var_type, defaults=settings) F2Tch||_||_||_||_||_||_||_dSrk)r%r?rvr=r}rri)rXr%r?rvr=r}rris r __init__zEstimatorSettings.__init__s9""  *, r"N)FFrrTFr)rrrrrr_r"r r3r3Rs=--\JLBDr"r3ceZdZdZdZdZdS) LeaveOneOuta Generator to give leave-one-out views on X. Parameters ---------- X : array_like 2-D array. Examples -------- >>> X = np.random.normal(0, 1, [10,2]) >>> loo = LeaveOneOut(X) >>> for x in loo: ... print x Notes ----- A little lighter weight than sklearn LOO. We do not need test index. Also passes views on X, not the index. c8tj||_dSrk)rrTrrs r rzLeaveOneOut.__init__sAr"c#K|j}tj|\}}t|D]0}tj|t }d||<||ddfV1dS)N)dtypeF)rrrrwonesbool)rXrrprFrgindexs r __iter__zLeaveOneOut.__iter__su Fx{{ ft  AGD---EE!HE111H+      r"N)rrrrrrr_r"r rrs<(r"rctjd|D}tjd|D}tjd|D}|||fS)Ncg|]}|dk Srcr_rfrcs r rhz!_get_type_pos..s333QS333r"cg|]}|dk S)or_rs r rhz!_get_type_pos..s222AqCx222r"cg|]}|dk S)ur_rs r rhz!_get_type_pos..s444ac444r")rarray)r-ix_contr;r<s r rurusih33(33344G X22222 3 3Fx44844455H FH $$r"ctj|}|jdkrtj|}|jdkr |dkrd}n~|jdkr|dkrt |}n]tj|d|kr%tj|d|kr|j}tj|d}tj|||f}|S)z> Returns an array of shape (nobs, k_vars) for use with `gpke`.rr)rrTndimsqueezeryrTreshape)datrFrps r r5r5s *S//C x!||joo x1}}! Q6Q;;3xx 8C== v % %"(3--*:f*D*D%Cx}}Q *S4. ) )C Jr"r rrct|||}tj|j} t |D]B\} } t || } | || |dd| f|| | dd| f<Ctjd|D} | dtj|| z }|r|dS|S)a[ Returns the non-normalized Generalized Product Kernel Estimator Parameters ---------- bw : 1-D ndarray The user-specified bandwidth parameters. data : 1D or 2-D ndarray The training data. data_predict : 1-D ndarray The evaluation points at which the kernel estimation is performed. var_type : str, optional The variable type (continuous, ordered, unordered). ckertype : str, optional The kernel used for the continuous variables. okertype : str, optional The kernel used for the ordered discrete variables. ukertype : str, optional The kernel used for the unordered discrete variables. tosum : bool, optional Whether or not to sum the calculated array of densities. Default is True. Returns ------- dens : array_like The generalized product kernel density estimator. Notes ----- The formula for the multivariate kernel estimator for the pdf is: .. math:: f(x)=\frac{1}{nh_{1}...h_{q}}\sum_{i=1}^ {n}K\left(\frac{X_{i}-x}{h}\right) where .. math:: K\left(\frac{X_{i}-x}{h}\right) = k\left( \frac{X_{i1}-x_{1}}{h_{1}}\right)\times k\left( \frac{X_{i2}-x_{2}}{h_{2}}\right)\times...\times k\left(\frac{X_{iq}-x_{q}}{h_{q}}\right) )rcrrNcg|]}|dk Srr_rs r rhzgpke..s888!Q#X888r"rrr) dictrrzr enumerate kernel_funcrprodsum)r&r data_predictr-ckertypeokertypeukertypetosumkertypesKvaliivtypefunc iscontinuousdenss r gpkersXh(h777H 8DJ  Dx((BB E8E?+d2b642; R0@AAQQQU 888x88899L 99!9  rwr,'788 8D xxQx r")r rrT)$rrqnumpyrscipyrscipy.stats.mstatsrrlr{ ImportErrorrr wang_ryzinaitchison_aitkenr r wang_ryzin_reggaussian_convolutionwang_ryzin_convolutionaitchison_aitken_convolutionraitchison_aitken_cdfwang_ryzin_cdfrrrr!rJrLr3rrur5rr_r"r rs ))))))MMMJJJJJdW/#*#;#,*1*F#*#9%,%A)0)G/6/S ' 4'.'C!(!7%0"? , , , 3$3$3$ll"l"l"l"l"l"l"l"^77777777tD%%%&5?AE>>>>>>s ))