0PhLXdZddlZddlmZddlmZddlZddlm Z ddl m Z m Z m Z mZmZddlmZdd lmZdd lmZdd lmZmZdd lmZmZmZmZmZdd lm Z m!Z!m"Z"ddl#m$Z$ddl%m&Z&m'Z'ddl(m)Z)m*Z*ddl+m,Z,m-Z-m.Z.m/Z/m0Z0ddl1m2Z2m3Z3dZ4Gdde2e e Z5Gdde5Z6dS)z1Recursive feature elimination for feature rankingN)deepcopy)Integral)effective_n_jobs) BaseEstimatorMetaEstimatorMixin _fit_contextclone is_classifier) get_scorer)check_cv_score)Bunchmetadata_routing)MetadataRouter MethodMapping_raise_for_params_routing_enabledprocess_routing) HasMethodsInterval RealNotInt)get_tags) _safe_split available_if)Paralleldelayed)_check_method_params_deprecate_positional_args_estimator_hascheck_is_fitted validate_data) SelectorMixin_get_feature_importancesc t||||\}} t|||||\ t||jj|} t||jj| |j||  fdfi| |j|jfS)zM Return the score and n_features per step for a fit across one fold. )paramsindices)Xr(r)c@t|dd|fS)N) score_paramsr) estimatorfeaturesX_testr,scorery_tests ^/var/www/html/test/jupyter/venv/lib/python3.11/site-packages/sklearn/feature_selection/_rfe.pyz!_rfe_single_fit..7s2F  111h;   % % % % ) rrr-fitr0score_fit step_scores_step_n_features_)rfer-r*ytraintestr0 routed_paramsX_trainy_train fit_paramsr/r,r1s ` @@@r2_rfe_single_fitrB's#9aE::GW Aq$>>NFF% -)-uJ( M(.L CH               S1 11r4c jeZdZUdZedggdeedddeedddgeedddeedddgd gee gd Z e e d <dddd d dZ edZedZeddZd"dZeeddZeeddZdZeeddZeeddZeeddZfd Zd!ZxZS)#RFEarFeature ranking with recursive feature elimination. Given an external estimator that assigns weights to features (e.g., the coefficients of a linear model), the goal of recursive feature elimination (RFE) is to select features by recursively considering smaller and smaller sets of features. First, the estimator is trained on the initial set of features and the importance of each feature is obtained either through any specific attribute or callable. Then, the least important features are pruned from current set of features. That procedure is recursively repeated on the pruned set until the desired number of features to select is eventually reached. Read more in the :ref:`User Guide `. Parameters ---------- estimator : ``Estimator`` instance A supervised learning estimator with a ``fit`` method that provides information about feature importance (e.g. `coef_`, `feature_importances_`). n_features_to_select : int or float, default=None The number of features to select. If `None`, half of the features are selected. If integer, the parameter is the absolute number of features to select. If float between 0 and 1, it is the fraction of features to select. .. versionchanged:: 0.24 Added float values for fractions. step : int or float, default=1 If greater than or equal to 1, then ``step`` corresponds to the (integer) number of features to remove at each iteration. If within (0.0, 1.0), then ``step`` corresponds to the percentage (rounded down) of features to remove at each iteration. verbose : int, default=0 Controls verbosity of output. importance_getter : str or callable, default='auto' If 'auto', uses the feature importance either through a `coef_` or `feature_importances_` attributes of estimator. Also accepts a string that specifies an attribute name/path for extracting feature importance (implemented with `attrgetter`). For example, give `regressor_.coef_` in case of :class:`~sklearn.compose.TransformedTargetRegressor` or `named_steps.clf.feature_importances_` in case of class:`~sklearn.pipeline.Pipeline` with its last step named `clf`. If `callable`, overrides the default feature importance getter. The callable is passed with the fitted estimator and it should return importance for each feature. .. versionadded:: 0.24 Attributes ---------- classes_ : ndarray of shape (n_classes,) The classes labels. Only available when `estimator` is a classifier. estimator_ : ``Estimator`` instance The fitted estimator used to select features. n_features_ : int The number of selected features. n_features_in_ : int Number of features seen during :term:`fit`. Only defined if the underlying estimator exposes such an attribute when fit. .. versionadded:: 0.24 feature_names_in_ : ndarray of shape (`n_features_in_`,) Names of features seen during :term:`fit`. Defined only when `X` has feature names that are all strings. .. versionadded:: 1.0 ranking_ : ndarray of shape (n_features,) The feature ranking, such that ``ranking_[i]`` corresponds to the ranking position of the i-th feature. Selected (i.e., estimated best) features are assigned rank 1. support_ : ndarray of shape (n_features,) The mask of selected features. See Also -------- RFECV : Recursive feature elimination with built-in cross-validated selection of the best number of features. SelectFromModel : Feature selection based on thresholds of importance weights. SequentialFeatureSelector : Sequential cross-validation based feature selection. Does not rely on importance weights. Notes ----- Allows NaN/Inf in the input if the underlying estimator does as well. References ---------- .. [1] Guyon, I., Weston, J., Barnhill, S., & Vapnik, V., "Gene selection for cancer classification using support vector machines", Mach. Learn., 46(1-3), 389--422, 2002. Examples -------- The following example shows how to retrieve the 5 most informative features in the Friedman #1 dataset. >>> from sklearn.datasets import make_friedman1 >>> from sklearn.feature_selection import RFE >>> from sklearn.svm import SVR >>> X, y = make_friedman1(n_samples=50, n_features=10, random_state=0) >>> estimator = SVR(kernel="linear") >>> selector = RFE(estimator, n_features_to_select=5, step=1) >>> selector = selector.fit(X, y) >>> selector.support_ array([ True, True, True, True, True, False, False, False, False, False]) >>> selector.ranking_ array([1, 1, 1, 1, 1, 6, 4, 3, 2, 5]) r5Nrr$rightclosedneitherverbose)r-n_features_to_selectsteprIimportance_getter_parameter_constraintsauto)rJrKrIrLcL||_||_||_||_||_dSNr-rJrKrLrI)selfr-rJrKrIrLs r2__init__z RFE.__init__s-#$8! !2 r4c|jjSrP)r-_estimator_typerRs r2rUzRFE._estimator_types ~--r4c|jjS)zClasses labels available when `estimator` is a classifier. Returns ------- ndarray of shape (n_classes,) ) estimator_classes_rVs r2rYz RFE.classes_s''r4Fprefer_skip_nested_validationc trt|dfi|}ntt|}|j||fi|jjS)abFit the RFE model and then the underlying estimator on the selected features. Parameters ---------- X : {array-like, sparse matrix} of shape (n_samples, n_features) The training input samples. y : array-like of shape (n_samples,) The target values. **fit_params : dict - If `enable_metadata_routing=False` (default): Parameters directly passed to the ``fit`` method of the underlying estimator. - If `enable_metadata_routing=True`: Parameters safely routed to the ``fit`` method of the underlying estimator. .. versionchanged:: 1.6 See :ref:`Metadata Routing User Guide ` for more details. Returns ------- self : object Fitted estimator. r5r5r-)rrrr7r-r5)rRr*r;rAr>s r2r5zRFE.fitsi>    C+D%FF:FFMM!Ej,A,A,ABBBMtyA==!8!<===r4c t|||dddd\}}|jd}|j|dz}n`t|jtr/|j}||kr!t jd|d|d tnt||jz}d |j cxkrd kr)nn&ttd|j |z}nt|j }tj |t }tj |t } |rg|_g|_tj||krutj||} t%|j} |jd kr$t+dtj|z| j|dd| f|fi|t/| |jd} tj| } tj| } t7|tj||z }|rK|jt;| |j|| | d|| | d|<| tj|xxdz cc<tj||kutj||} t%|j|_|jj|dd| f|fi||rP|jt;| |j||j| ||_ ||_!| |_"|S)NcscrFT accept_sparseensure_min_featuresensure_all_finite multi_outputr$zFound n_features_to_select= > n_features=C. There will be no feature selection and all features will be kept.gg?)dtyperz#Fitting estimator with %d features.square)transform_func)#r#shaperJ isinstancerwarningswarn UserWarningintrKmaxnponesboolr9r8sumaranger r-rIprintr5r&rLargsortravelminappendlen logical_notrX n_features_support_ranking_)rRr*r; step_scorerA n_featuresrJrKrrr.r- importancesranks thresholds r2r7zRFE._fits    !#   1WQZ  $ ,#-? 18 < < O#'#< #j00 O!5OO:OOO $'zD4M'M#N#N  S s1di*45566DDty>>D7:T2227:S111  #$&D ! "D fX!555y,,X6Hdn--I|a;bfX>N>NNOOO IM!AAAxK.! : :z : : :3&'K J{++EHUOOED"&"2"25I"IJJI  J%,,S]];;;!((Ix)H)HIII49HXe_ZiZ0 1 R^H-- . . .! 3 . . .AfX!555F9Z((2//AaaakNA<<<<<  L  ! ( (X 7 7 7   $ $ZZ%J%J K K K#<<>>     r4predictc t||dt|trt|dfi|}nt t i}|jj||fi|jjS)aReduce X to the selected features and predict using the estimator. Parameters ---------- X : array of shape [n_samples, n_features] The input samples. **predict_params : dict Parameters to route to the ``predict`` method of the underlying estimator. .. versionadded:: 1.6 Only available if `enable_metadata_routing=True`, which can be set by using ``sklearn.set_config(enable_metadata_routing=True)``. See :ref:`Metadata Routing User Guide ` for more details. Returns ------- y : array of shape [n_samples] The predicted target values. r)rr^) rr"rrrrXr transformr-)rRr*predict_paramsr>s r2rz RFE.predictts2 .$ :::    ?+D)NN~NNMM!E",=,=,=>>>M&t& NN1    !.!8!@   r4r6c t|trt|dfi|}ntt|}|jj|||fi|jjS)aReduce X to the selected features and return the score of the estimator. Parameters ---------- X : array of shape [n_samples, n_features] The input samples. y : array of shape [n_samples] The target values. **score_params : dict - If `enable_metadata_routing=False` (default): Parameters directly passed to the ``score`` method of the underlying estimator. - If `enable_metadata_routing=True`: Parameters safely routed to the `score` method of the underlying estimator. .. versionadded:: 1.0 .. versionchanged:: 1.6 See :ref:`Metadata Routing User Guide ` for more details. Returns ------- score : float Score of the underlying base estimator computed with the selected features returned by `rfe.transform(X)` and `y`. r6r6r^)r"rrrrXr6rr-)rRr*r;r,r>s r2r6z RFE.scores>     G+D'JJ\JJMM!E ,E,E,EFFFM$t$ NN1  q  $1$;$A   r4c.t||jSrP)r"rrVs r2_get_support_maskzRFE._get_support_masks}r4decision_functionczt||j||S)aCompute the decision function of ``X``. Parameters ---------- X : {array-like or sparse matrix} of shape (n_samples, n_features) The input samples. Internally, it will be converted to ``dtype=np.float32`` and if a sparse matrix is provided to a sparse ``csr_matrix``. Returns ------- score : array, shape = [n_samples, n_classes] or [n_samples] The decision function of the input samples. The order of the classes corresponds to that in the attribute :term:`classes_`. Regression and binary classification produce an array of shape [n_samples]. )r"rXrrrRr*s r2rzRFE.decision_functions4& 001B1BCCCr4 predict_probaczt||j||S)a5Predict class probabilities for X. Parameters ---------- X : {array-like or sparse matrix} of shape (n_samples, n_features) The input samples. Internally, it will be converted to ``dtype=np.float32`` and if a sparse matrix is provided to a sparse ``csr_matrix``. Returns ------- p : array of shape (n_samples, n_classes) The class probabilities of the input samples. The order of the classes corresponds to that in the attribute :term:`classes_`. )r"rXrrrs r2rzRFE.predict_probas4" ,,T^^A->->???r4predict_log_probaczt||j||S)aPredict class log-probabilities for X. Parameters ---------- X : array of shape [n_samples, n_features] The input samples. Returns ------- p : array of shape (n_samples, n_classes) The class log-probabilities of the input samples. The order of the classes corresponds to that in the attribute :term:`classes_`. )r"rXrrrs r2rzRFE.predict_log_probas4 001B1BCCCr4ct}t|j}|j|_t |j|_t |j|_|j d|j_|j d|j_d|j _ |j j |j _ |j j |j _ |S)NT)super__sklearn_tags__rr-estimator_typerclassifier_tagsregressor_tags poor_score target_tagsrequired input_tagssparse allow_nan)rRtagssub_estimator_tags __class__s r2rzRFE.__sklearn_tags__sww''))%dn550?'(:(JKK&'9'HII   +.2D +   *-1D  *$(!!3!>!E$6$A$K! r4ct|jj|jt dddddd}|S)jGet metadata routing of this object. Please check :ref:`User Guide ` on how the routing mechanism works. .. versionadded:: 1.6 Returns ------- routing : MetadataRouter A :class:`~sklearn.utils.metadata_routing.MetadataRouter` encapsulating routing information. ownerr5callercalleerr6r-method_mapping)rr__name__addr-rrRrouters r2get_metadata_routingzRFE.get_metadata_routingsu dn&=>>>BBn(?? SeS , , S )S 4 4 SS 0 0 C   r4rP)r __module__ __qualname____doc__rrrrstrcallablerMdict__annotations__rSpropertyrUrYr r5r7rr!rr6rrrrrr __classcell__)rs@r2rDrDDs||~!j%))*  HZAg 6 6 6 HXq$y 9 9 9! HXq$y 9 9 9 HZAi 8 8 8 ;!8_ $ $D   &"        ..X.((X(\&+ > >  >D\\\\|\..++,,! ! -,! F\..))**& & +*& P\..!45566DD76D*\..1122@@32@&\..!45566DD76D"     r4rDc eZdZUdZiejeedddgdgdee gdegdZe e d<e d d e jiZd d ddddd d dZededdddZdZdZdZdS)RFECVaqRecursive feature elimination with cross-validation to select features. The number of features selected is tuned automatically by fitting an :class:`RFE` selector on the different cross-validation splits (provided by the `cv` parameter). The performance of the :class:`RFE` selector are evaluated using `scorer` for different number of selected features and aggregated together. Finally, the scores are averaged across folds and the number of features selected is set to the number of features that maximize the cross-validation score. See glossary entry for :term:`cross-validation estimator`. Read more in the :ref:`User Guide `. Parameters ---------- estimator : ``Estimator`` instance A supervised learning estimator with a ``fit`` method that provides information about feature importance either through a ``coef_`` attribute or through a ``feature_importances_`` attribute. step : int or float, default=1 If greater than or equal to 1, then ``step`` corresponds to the (integer) number of features to remove at each iteration. If within (0.0, 1.0), then ``step`` corresponds to the percentage (rounded down) of features to remove at each iteration. Note that the last iteration may remove fewer than ``step`` features in order to reach ``min_features_to_select``. min_features_to_select : int, default=1 The minimum number of features to be selected. This number of features will always be scored, even if the difference between the original feature count and ``min_features_to_select`` isn't divisible by ``step``. .. versionadded:: 0.20 cv : int, cross-validation generator or an iterable, default=None Determines the cross-validation splitting strategy. Possible inputs for cv are: - None, to use the default 5-fold cross-validation, - integer, to specify the number of folds. - :term:`CV splitter`, - An iterable yielding (train, test) splits as arrays of indices. For integer/None inputs, if ``y`` is binary or multiclass, :class:`~sklearn.model_selection.StratifiedKFold` is used. If the estimator is not a classifier or if ``y`` is neither binary nor multiclass, :class:`~sklearn.model_selection.KFold` is used. Refer :ref:`User Guide ` for the various cross-validation strategies that can be used here. .. versionchanged:: 0.22 ``cv`` default value of None changed from 3-fold to 5-fold. scoring : str, callable or None, default=None A string (see :ref:`scoring_parameter`) or a scorer callable object / function with signature ``scorer(estimator, X, y)``. verbose : int, default=0 Controls verbosity of output. n_jobs : int or None, default=None Number of cores to run in parallel while fitting across folds. ``None`` means 1 unless in a :obj:`joblib.parallel_backend` context. ``-1`` means using all processors. See :term:`Glossary ` for more details. .. versionadded:: 0.18 importance_getter : str or callable, default='auto' If 'auto', uses the feature importance either through a `coef_` or `feature_importances_` attributes of estimator. Also accepts a string that specifies an attribute name/path for extracting feature importance. For example, give `regressor_.coef_` in case of :class:`~sklearn.compose.TransformedTargetRegressor` or `named_steps.clf.feature_importances_` in case of :class:`~sklearn.pipeline.Pipeline` with its last step named `clf`. If `callable`, overrides the default feature importance getter. The callable is passed with the fitted estimator and it should return importance for each feature. .. versionadded:: 0.24 Attributes ---------- classes_ : ndarray of shape (n_classes,) The classes labels. Only available when `estimator` is a classifier. estimator_ : ``Estimator`` instance The fitted estimator used to select features. cv_results_ : dict of ndarrays All arrays (values of the dictionary) are sorted in ascending order by the number of features used (i.e., the first element of the array represents the models that used the least number of features, while the last element represents the models that used all available features). .. versionadded:: 1.0 This dictionary contains the following keys: split(k)_test_score : ndarray of shape (n_subsets_of_features,) The cross-validation scores across (k)th fold. mean_test_score : ndarray of shape (n_subsets_of_features,) Mean of scores over the folds. std_test_score : ndarray of shape (n_subsets_of_features,) Standard deviation of scores over the folds. n_features : ndarray of shape (n_subsets_of_features,) Number of features used at each step. .. versionadded:: 1.5 n_features_ : int The number of selected features with cross-validation. n_features_in_ : int Number of features seen during :term:`fit`. Only defined if the underlying estimator exposes such an attribute when fit. .. versionadded:: 0.24 feature_names_in_ : ndarray of shape (`n_features_in_`,) Names of features seen during :term:`fit`. Defined only when `X` has feature names that are all strings. .. versionadded:: 1.0 ranking_ : narray of shape (n_features,) The feature ranking, such that `ranking_[i]` corresponds to the ranking position of the i-th feature. Selected (i.e., estimated best) features are assigned rank 1. support_ : ndarray of shape (n_features,) The mask of selected features. See Also -------- RFE : Recursive feature elimination. Notes ----- The size of all values in ``cv_results_`` is equal to ``ceil((n_features - min_features_to_select) / step) + 1``, where step is the number of features removed at each iteration. Allows NaN/Inf in the input if the underlying estimator does as well. References ---------- .. [1] Guyon, I., Weston, J., Barnhill, S., & Vapnik, V., "Gene selection for cancer classification using support vector machines", Mach. Learn., 46(1-3), 389--422, 2002. Examples -------- The following example shows how to retrieve the a-priori not known 5 informative features in the Friedman #1 dataset. >>> from sklearn.datasets import make_friedman1 >>> from sklearn.feature_selection import RFECV >>> from sklearn.svm import SVR >>> X, y = make_friedman1(n_samples=50, n_features=10, random_state=0) >>> estimator = SVR(kernel="linear") >>> selector = RFECV(estimator, step=1, cv=5) >>> selector = selector.fit(X, y) >>> selector.support_ array([ True, True, True, True, True, False, False, False, False, False]) >>> selector.ranking_ array([1, 1, 1, 1, 1, 6, 4, 3, 2, 5]) rNrHrF cv_object)min_features_to_selectcvscoringn_jobsrMrJgroupsr$rN)rKrrrrIrrLcv||_||_||_||_||_||_||_||_dSrP)r-rKrLrrrIrr) rRr-rKrrrrIrrLs r2rSzRFECV.__init__sD# !2   &<###r4z1.8)versionFrZ)rc t|dtdddd\tr(||d|it dfi|n>t t i t d|i t i  t jtj  } j d}j |kr&tjdj d|dtt!j t#j |jjjt+jdkrt.t0c}n)t3j}t5t0|fd|jfijjD}t;|\} } t=j| dddd} t=j| } t=j | dddd} | t=j!| } t!j | jjjj"fij j"j#_#j$_$j%_%tMj _'j'j"(fij j"| dddddft=j)dt=j*ddfdtW| j dDd| i_,S)aaFit the RFE model and automatically tune the number of selected features. Parameters ---------- X : {array-like, sparse matrix} of shape (n_samples, n_features) Training vector, where `n_samples` is the number of samples and `n_features` is the total number of features. y : array-like of shape (n_samples,) Target values (integers for classification, real numbers for regression). groups : array-like of shape (n_samples,) or None, default=None Group labels for the samples used while splitting the dataset into train/test set. Only used in conjunction with a "Group" :term:`cv` instance (e.g., :class:`~sklearn.model_selection.GroupKFold`). .. versionadded:: 0.20 **params : dict of str -> object Parameters passed to the ``fit`` method of the estimator, the scorer, and the CV splitter. .. versionadded:: 1.6 Only available if `enable_metadata_routing=True`, which can be set by using ``sklearn.set_config(enable_metadata_routing=True)``. See :ref:`Metadata Routing User Guide ` for more details. Returns ------- self : object Fitted estimator. r5csrrFTraNrr])splitr)r-splitterr0) classifierr$zFound min_features_to_select=rfrg)r-rJrLrKrI)rc 3fK|]+\}}tj ||V,dSrP)r r-) .0r<r=r*funcr:r>r0rRr;s r2 zRFECV.fit..is[# # t DsT^Q5$ V V# # # # # # r4r)axisrQ)mean_test_scorestd_test_scorec*i|]}d|d|S)r _test_score)ri scores_revs r2 zRFECV.fit..s+UUU%q%%%z!}UUUr4r)-rr#rupdaterrr rr r- _get_scorerrkrrmrnrorDrzrLrKrIrrlistrBrrrrziprrarrayruargmaxr5rr~rr rX _transformmeanstdrange cv_results_)rRr*r;rr(rrparallelscores_featuresscoresstep_n_featuresstep_n_features_revscores_sum_revrJrr:r>r0rs``` @@@@@r2r5z RFECV.fitsR &$...   !#   1    ! x0111+D%BB6BBMM!B---h%78882Mdgq]4>-J-J K K K!!##WQZ  & 3 3 M-D4O--!---    n!$T%@*!M!M"4L    ( DK ( (A - -!?NHddt{333H?++D"(# # # # # # # # # # 'rx1MM 0F0LMM# # #   #&"7 hq'9::44R4@&!!Q///"5229^3L3LMn!5"4L     144 /3444  ?  //DOOA..QQ]5L5PQQQAAAtttG_ !wz::: fZa888  VUUUeFLQRO>T>TUUU  -     r4c t||d|}trt|dfi|}n#t }t i|_||||fi|jjS)a<Score using the `scoring` option on the given test data and labels. Parameters ---------- X : array-like of shape (n_samples, n_features) Test samples. y : array-like of shape (n_samples,) True labels for X. **score_params : dict Parameters to pass to the `score` method of the underlying scorer. .. versionadded:: 1.6 Only available if `enable_metadata_routing=True`, which can be set by using ``sklearn.set_config(enable_metadata_routing=True)``. See :ref:`Metadata Routing User Guide ` for more details. Returns ------- score : float Score of self.predict(X) w.r.t. y defined by `scoring`. r6r)rrrrrr0r6)rRr*r;r,rr>s r2r6z RFECV.scores4 ,g666""$$    3+D'JJ\JJMM!GGM#(r???M wtQ@@]%9%?@@@r4ct|jj}||jt dd|t |jt dd||t dddd|S) rrr5rrr)rrr6)r0r) rrrrr-rr rrrs r2rzRFECV.get_metadata_routings dn&=>>> n(??..eE.JJ      dg&&(??../      ##%%(?? SgS . . SS 0 0     r4cn|jt|jrdnd}n|j}t|S)Naccuracyr2)rr r-r )rRrs r2rzRFECV._get_scorers9 < $1$.$A$AKjjtGGlG'"""r4)rrrrrDrMrrrrrrpoprUNUSED_RFECV__metadata_request__fitrSr r r5r6rrrr4r2rr(snuun$ $$#+8Hai#P#P#P"Qm#x(" $$$D5666')9)@A   =====, ...\&+#'KKKK /. KZ"A"A"AH!!!F#####r4r)7rrmcopyrnumbersrnumpyrrjoblibrbaserrr r r metricsr model_selectionr model_selection._validationrutilsrrutils._metadata_requestsrrrrrutils._param_validationrrr utils._tagsrutils.metaestimatorsrrutils.parallelrrutils.validationrr r!r"r#_baser%r&rBrDrrr4r2rs^87######XXXXXXXXXXXXXX &&&&&&000000++++++++GFFFFFFFFF""""""<<<<<<<<........;:::::::222:aaaaa-+]aaaHv#v#v#v#v#Cv#v#v#v#v#r4