-ib SSKJr SSKJrJrJr SSKJrJr SSK r SSK J r J r SSKJr SSKJr SS KJrJr SS KJr SS KJrJr S S KJrJr S SKJr S SKJ r J!r! SS/r""SS5r#Sr$"SS\5r%"SS\%5r&"SS\%5r'g))abstractmethod)ceilfloorlog)IntegralRealN) _fit_context is_classifier)get_scorer_names)resample)Interval StrOptions)check_classification_targets) _num_samples validate_data) ParameterGridParameterSampler) BaseSearchCV)_yields_constant_splitscheck_cvHalvingGridSearchCVHalvingRandomSearchCVc$\rSrSrSrSrSrSrg)_SubsampleMetaSplitterz8Splitter that subsamples a given fraction of the datasetc4XlX lX0lX@lgNbase_cvfractionsubsample_test random_state)selfr!r"r#r$s e/var/www/html/venv/lib/python3.13/site-packages/sklearn/model_selection/_search_successive_halving.py__init___SubsampleMetaSplitter.__init__s  ,(c +X# URR"X40UD6HupE[USUR[ UR [ U5-5S9nUR(a5[USUR[ UR [ U5-5S9nXE4v M g7f)NF)replacer$ n_samples)r!splitr r$intr"lenr#)r%Xykwargs train_idxtest_idxs r&r-_SubsampleMetaSplitter.split s#'<<#5#5a#Ef#E I !..dmmc)n<= I ""#!!%!2!2!$--#h-"?@  % %$FsB(B*)r!r"r$r#N)__name__ __module__ __qualname____firstlineno____doc__r'r-__static_attributes__r)r&rrsB) &r)rcLSUSUSUS45up4n[R"X2:H5nXFn[R"[R"U5[R"[R "U555n[R "XVX*S5$)Nc3N# UHn[R"U5v M g7fr)npasarray).0as r& _top_k..4s!*QA 1 Qs#%itermean_test_scoreparams)r? flatnonzerorollargsort count_nonzeroisnanarray) resultskitr iterationrFrG iter_indicesscoressorted_indicess r&_top_krU2s*&/7+<#=wx?PQ*&I>>)"23L  *FWWRZZ/1A1A"((6BR1STN 88F()<= >>r)c^\rSrSr%Sr0\R E\"\"\ "555\ S/S/\ "\ SSSS9\"S15/\ "\ SSSS9\"S S 15/\ /\ "\SSSS9/S /S .Er\\S '\R#S5 SSSSSS\R&SSS SSSS. U4SjjrSr\S5r\"SS9SU4Sjj5rSr\S5rSrU=r$)BaseSuccessiveHalvingAzImplements successive halving. Ref: Almost optimal exploration in multi-armed bandits, ICML 13 Zohar Karnin, Tomer Koren, Oren Somekh Nr$rneitherclosedautoexhaustsmallestboolean)scoringr$ max_resources min_resourcesresourcefactoraggressive_elimination_parameter_constraints pre_dispatchTr,F) r`n_jobsrefitcvverboser$ error_scorereturn_train_scorerarbrcrdrec v>[TU]UUUUUUUU S9 XplXlXlXlXlXlg)N)r`rjrkrlrmrnro)superr'r$rarcrdrbre)r% estimatorr`rjrkrlrmr$rnrorarbrcrdre __class__s r&r'BaseSuccessiveHalving.__init__\sS$  #1  )*  *&<#r)c[UR5(d [S5eURS:waaURURR 5;a9[SURSURR R35e[U[5(a/URURs=:XaS:XaO O [S5eURUl URS;aURS:XaURR"X40UD6nSnXE-Ul [UR5(aN[US US 9n[!U5 ["R$"U5R&S nU=RU-sl OS Ul UR(UlUR*S :Xa+URS:Xd [S5e[-U5UlURUR*:a&[SURSUR*S35eURS :Xa[SURS35eg)Nz{The cv parameter must yield consistent folds across calls to split(). Set its random_state to an int, or set shuffle=False.r,zCannot use resource=z% which is not supported by estimator r]z?n_candidates and min_resources cannot be both set to 'exhaust'.)r^r]r no_validation)r0r1rrr\z:resource can only be 'n_samples' when max_resources='auto'zmin_resources_=z is greater than max_resources_=.z+: you might have passed an empty dataset X.)r_checked_cv_orig ValueErrorrcrr get_paramsrsr6 isinstancerrb n_candidatesmin_resources_ get_n_splitsr rrr?uniqueshaperamax_resources_r)r%r0r1 split_paramsn_splits magic_factor n_classess r&_check_input_parameters-BaseSuccessiveHalving._check_input_parameterss8't'<'<==!  MM[ ( T^^%>%>%@@&t}}o6 $ 8 8 A ABD  d1 2 2!!T%6%6C)C!U#00   "9 9}} +00==aSlS &.&=# 00%doCA03 " ! 2 21 5I''94'&'##00   & (==K/ P#/q/D    !4!4 4!$"5"5!67''+':':&;1>    ! #!$"5"5!67&&  $r)c[R"US5n[R"USU:H5nUSUn[R"U5R 5(aSnXF$[R "U5nXF$)a)Custom refit callable to return the index of the best candidate. We want the best candidate out of the last iteration. By default BaseSearchCV would return the best candidate out of all iterations. Currently, we only support for a single metric thus `refit` and `refit_metric` are not required. rErFr)r?maxrHrLall nanargmax)rk refit_metricrN last_iterlast_iter_indices test_scoresbest_idxs r&_select_best_index(BaseSuccessiveHalving._select_best_indexsFF76?+ NN76?i+GH/01BC  88K $ $ & &H!**||K0H **r))prefer_skip_nested_validationc T>[URU[UR5S9UlUR U5nUR XURRS9 [U5Ul [TU]0"U4SU0UD6 URSURUlU$)a%Run fit with all sets of parameters. Parameters ---------- X : array-like, shape (n_samples, n_features) Training vector, where `n_samples` is the number of samples and `n_features` is the number of features. y : array-like, shape (n_samples,) or (n_samples, n_output), optional Target relative to X for classification or regression; None for unsupervised learning. **params : dict of string -> object Parameters passed to the ``fit`` method of the estimator. Returns ------- self : object Instance of fitted estimator. ) classifier)r0r1rr1rF)rrlr rrrx_get_routed_params_for_fitrsplitterr-r_n_samples_origrqfit cv_results_ best_index_ best_score_)r%r0r1rG routed_paramsrss r&rBaseSuccessiveHalving.fits4!) GGQ=#@! 77?  $$=#9#9#?#? %  ,A  A%%f% ++,=>t?O?OP r)cR^TR5nTRS:wa3[U4SjU55(a[STRS35eS[ [ [ U5TR55-nTRS:Xa:US- n[TRTRTRU--5Tl S[ [ TRTR-TR55-nTR(aUnO [XS5nTR(a[SU35 [SU35 [S U35 [S TR35 [S TR35 [S TR35 [S TR35 /Tl/Tl[%U5GHnUnTR(a[SXs- U-5n['TRU-TR-5n [U TR5n TR R)U 5 [ U5n TR"R)U 5 TR(a5[S5 [SU35 [SU 35 [SU 35 TRS:Xa.[+TR,U TR.- STR0S9n OBUV s/sHoR35PM nn UHn XTR'M TR,n U/U -U /U -S.nU"X+US9n[5U TR- 5n[7UUU5nGM [ U5TlUTlUTlUTlgs sn f)Nr,c3B># UHnTRU;v M g7fr)rc)rA candidater%s r&rC4BaseSuccessiveHalving._run_search..s0 8H9DMMY &8HszCannot use parameter z= as the resource since it is part of the searched parameters.rr]zn_iterations: zn_required_iterations: zn_possible_iterations: zmin_resources_: zmax_resources_: zaggressive_elimination: zfactor: rz ----------ziter: zn_candidates: z n_resources: Tr )rE n_resources) more_results) _generate_candidate_paramsrcanyryrrr/rdrbrr}rreminrmprint n_resources_ n_candidates_ranger.appendrrxrr$copyrrUn_remaining_candidates_n_required_iterations_n_possible_iterations_ n_iterations_)r%evaluate_candidatescandidate_paramsn_required_iterationslast_iterationn_possible_iterations n_iterationsrPpowerrr|rlcrrrNn_candidates_to_keeps` r& _run_search!BaseSuccessiveHalving._run_searchs::< ==K 'C0 8H0 - - ' 799 !"E#c2B.CT[[*Q$R R    *3Q6N"%####t{{N'BB#D !"E ##t':'::DKK H% !   & &0L4LL << N<.1 2 +,A+BC D +,A+BC D $T%8%8$9: ; $T%8%8$9: ; ,T-H-H,IJ K HT[[M* +&CE** As:=RRSdkk5043F3FFGKk4+>+>?K    $ $[ 1/0L    % %l 3||hsen%|n56 k]34}} ++ 11(4+?+??#'!%!2!2 7G#G6FFFH6F #G!1I/:dmm,"2** , +}|;L * <G$( t{{(B#C %g/CSI e'h(++;'<$&;#&;#))$HsN$cgrr<r%s r&r0BaseSuccessiveHalving._generate_candidate_paramsqs r))rxrrerrdrarrbr}rrrrrrr$rcr)r6r7r8r9r:rrfrsetr callablerrstrrdict__annotations__popr?nanr'r staticmethodrr rrrrr; __classcell__rss@r&rWrWAs@$  - -$s#3#5674H'( Xq$y 9 x  Xq$y 9  :. / ED!T)<=#,+$D"~.  FF$!"="=H?B++,&+& &Pk*Z  r)rWc^\rSrSr%Sr0\R ES\\/0Er\\ S'SSSSS S S S \ RS S S S S. U4Sjjr Sr SrU=r$)riva0Search over specified parameter values with successive halving. The search strategy starts evaluating all the candidates with a small amount of resources and iteratively selects the best candidates, using more and more resources. Read more in the :ref:`User guide `. .. note:: This estimator is still **experimental** for now: the predictions and the API might change without any deprecation cycle. To use it, you need to explicitly import ``enable_halving_search_cv``:: >>> # explicitly require this experimental feature >>> from sklearn.experimental import enable_halving_search_cv # noqa >>> # now you can import normally from model_selection >>> from sklearn.model_selection import HalvingGridSearchCV Parameters ---------- estimator : estimator object This is assumed to implement the scikit-learn estimator interface. Either estimator needs to provide a ``score`` function, or ``scoring`` must be passed. param_grid : dict or list of dictionaries Dictionary with parameters names (string) as keys and lists of parameter settings to try as values, or a list of such dictionaries, in which case the grids spanned by each dictionary in the list are explored. This enables searching over any sequence of parameter settings. factor : int or float, default=3 The 'halving' parameter, which determines the proportion of candidates that are selected for each subsequent iteration. For example, ``factor=3`` means that only one third of the candidates are selected. resource : ``'n_samples'`` or str, default='n_samples' Defines the resource that increases with each iteration. By default, the resource is the number of samples. It can also be set to any parameter of the base estimator that accepts positive integer values, e.g. 'n_iterations' or 'n_estimators' for a gradient boosting estimator. In this case ``max_resources`` cannot be 'auto' and must be set explicitly. max_resources : int, default='auto' The maximum amount of resource that any candidate is allowed to use for a given iteration. By default, this is set to ``n_samples`` when ``resource='n_samples'`` (default), else an error is raised. min_resources : {'exhaust', 'smallest'} or int, default='exhaust' The minimum amount of resource that any candidate is allowed to use for a given iteration. Equivalently, this defines the amount of resources `r0` that are allocated for each candidate at the first iteration. - 'smallest' is a heuristic that sets `r0` to a small value: - ``n_splits * 2`` when ``resource='n_samples'`` for a regression problem - ``n_classes * n_splits * 2`` when ``resource='n_samples'`` for a classification problem - ``1`` when ``resource != 'n_samples'`` - 'exhaust' will set `r0` such that the **last** iteration uses as much resources as possible. Namely, the last iteration will use the highest value smaller than ``max_resources`` that is a multiple of both ``min_resources`` and ``factor``. In general, using 'exhaust' leads to a more accurate estimator, but is slightly more time consuming. Note that the amount of resources used at each iteration is always a multiple of ``min_resources``. aggressive_elimination : bool, default=False This is only relevant in cases where there isn't enough resources to reduce the remaining candidates to at most `factor` after the last iteration. If ``True``, then the search process will 'replay' the first iteration for as long as needed until the number of candidates is small enough. This is ``False`` by default, which means that the last iteration may evaluate more than ``factor`` candidates. See :ref:`aggressive_elimination` for more details. cv : int, cross-validation generator or iterable, default=5 Determines the cross-validation splitting strategy. Possible inputs for cv are: - integer, to specify the number of folds in a `(Stratified)KFold`, - :term:`CV splitter`, - An iterable yielding (train, test) splits as arrays of indices. For integer/None inputs, if the estimator is a classifier and ``y`` is either binary or multiclass, :class:`StratifiedKFold` is used. In all other cases, :class:`KFold` is used. These splitters are instantiated with `shuffle=False` so the splits will be the same across calls. Refer :ref:`User Guide ` for the various cross-validation strategies that can be used here. .. note:: Due to implementation details, the folds produced by `cv` must be the same across multiple calls to `cv.split()`. For built-in `scikit-learn` iterators, this can be achieved by deactivating shuffling (`shuffle=False`), or by setting the `cv`'s `random_state` parameter to an integer. scoring : str or callable, default=None Scoring method to use to evaluate the predictions on the test set. - str: see :ref:`scoring_string_names` for options. - callable: a scorer callable object (e.g., function) with signature ``scorer(estimator, X, y)``. See :ref:`scoring_callable` for details. - `None`: the `estimator`'s :ref:`default evaluation criterion ` is used. refit : bool or callable, default=True Refit an estimator using the best found parameters on the whole dataset. Where there are considerations other than maximum score in choosing a best estimator, ``refit`` can be set to a function which returns the selected ``best_index_`` given ``cv_results_``. In that case, the ``best_estimator_`` and ``best_params_`` will be set according to the returned ``best_index_`` while the ``best_score_`` attribute will not be available. The refitted estimator is made available at the ``best_estimator_`` attribute and permits using ``predict`` directly on this ``HalvingGridSearchCV`` instance. See :ref:`this example ` for an example of how to use ``refit=callable`` to balance model complexity and cross-validated score. error_score : 'raise' or numeric Value to assign to the score if an error occurs in estimator fitting. If set to 'raise', the error is raised. If a numeric value is given, FitFailedWarning is raised. This parameter does not affect the refit step, which will always raise the error. Default is ``np.nan``. return_train_score : bool, default=False If ``False``, the ``cv_results_`` attribute will not include training scores. Computing training scores is used to get insights on how different parameter settings impact the overfitting/underfitting trade-off. However computing the scores on the training set can be computationally expensive and is not strictly required to select the parameters that yield the best generalization performance. random_state : int, RandomState instance or None, default=None Pseudo random number generator state used for subsampling the dataset when `resources != 'n_samples'`. Ignored otherwise. Pass an int for reproducible output across multiple function calls. See :term:`Glossary `. n_jobs : int or None, default=None Number of jobs to run in parallel. ``None`` means 1 unless in a :obj:`joblib.parallel_backend` context. ``-1`` means using all processors. See :term:`Glossary ` for more details. verbose : int Controls the verbosity: the higher, the more messages. Attributes ---------- n_resources_ : list of int The amount of resources used at each iteration. n_candidates_ : list of int The number of candidate parameters that were evaluated at each iteration. n_remaining_candidates_ : int The number of candidate parameters that are left after the last iteration. It corresponds to `ceil(n_candidates[-1] / factor)` max_resources_ : int The maximum number of resources that any candidate is allowed to use for a given iteration. Note that since the number of resources used at each iteration must be a multiple of ``min_resources_``, the actual number of resources used at the last iteration may be smaller than ``max_resources_``. min_resources_ : int The amount of resources that are allocated for each candidate at the first iteration. n_iterations_ : int The actual number of iterations that were run. This is equal to ``n_required_iterations_`` if ``aggressive_elimination`` is ``True``. Else, this is equal to ``min(n_possible_iterations_, n_required_iterations_)``. n_possible_iterations_ : int The number of iterations that are possible starting with ``min_resources_`` resources and without exceeding ``max_resources_``. n_required_iterations_ : int The number of iterations that are required to end up with less than ``factor`` candidates at the last iteration, starting with ``min_resources_`` resources. This will be smaller than ``n_possible_iterations_`` when there isn't enough resources. cv_results_ : dict of numpy (masked) ndarrays A dict with keys as column headers and values as columns, that can be imported into a pandas ``DataFrame``. It contains lots of information for analysing the results of a search. Please refer to the :ref:`User guide` for details. best_estimator_ : estimator or dict Estimator that was chosen by the search, i.e. estimator which gave highest score (or smallest loss if specified) on the left out data. Not available if ``refit=False``. best_score_ : float Mean cross-validated score of the best_estimator. best_params_ : dict Parameter setting that gave the best results on the hold out data. best_index_ : int The index (of the ``cv_results_`` arrays) which corresponds to the best candidate parameter setting. The dict at ``search.cv_results_['params'][search.best_index_]`` gives the parameter setting for the best model, that gives the highest mean score (``search.best_score_``). scorer_ : function or a dict Scorer function used on the held out data to choose the best parameters for the model. n_splits_ : int The number of cross-validation splits (folds/iterations). refit_time_ : float Seconds used for refitting the best model on the whole dataset. This is present only if ``refit`` is not False. multimetric_ : bool Whether or not the scorers compute several metrics. classes_ : ndarray of shape (n_classes,) The classes labels. This is present only if ``refit`` is specified and the underlying estimator is a classifier. n_features_in_ : int Number of features seen during :term:`fit`. Only defined if `best_estimator_` is defined (see the documentation for the `refit` parameter for more details) and that `best_estimator_` exposes `n_features_in_` when fit. .. versionadded:: 0.24 feature_names_in_ : ndarray of shape (`n_features_in_`,) Names of features seen during :term:`fit`. Only defined if `best_estimator_` is defined (see the documentation for the `refit` parameter for more details) and that `best_estimator_` exposes `feature_names_in_` when fit. .. versionadded:: 1.0 See Also -------- :class:`HalvingRandomSearchCV`: Random search over a set of parameters using successive halving. Notes ----- The parameters selected are those that maximize the score of the held-out data, according to the scoring parameter. All parameter combinations scored with a NaN will share the lowest rank. Examples -------- >>> from sklearn.datasets import load_iris >>> from sklearn.ensemble import RandomForestClassifier >>> from sklearn.experimental import enable_halving_search_cv # noqa >>> from sklearn.model_selection import HalvingGridSearchCV ... >>> X, y = load_iris(return_X_y=True) >>> clf = RandomForestClassifier(random_state=0) ... >>> param_grid = {"max_depth": [3, None], ... "min_samples_split": [5, 10]} >>> search = HalvingGridSearchCV(clf, param_grid, resource='n_estimators', ... max_resources=10, ... random_state=0).fit(X, y) >>> search.best_params_ # doctest: +SKIP {'max_depth': None, 'min_samples_split': 10, 'n_estimators': 9} param_gridrfrir,r\r]FrhNTr) rdrcrarbrerlr`rkrnror$rjrmc F>[TU]UU UU UUU U U UUUUUS9 X lgN) r`rjrkrmrlr$rnrorarcrdrbre)rqr'r)r%rrrrdrcrarbrerlr`rkrnror$rjrmrss r&r'HalvingGridSearchCV.__init__sH&  %#1''#9  %r)c,[UR5$r)rrrs r&r.HalvingGridSearchCV._generate_candidate_paramssT__--r))r)r6r7r8r9r:rWrfrlistrr?rr'rr;rrs@r&rrvswiV $  6 6$tTl$D$ FF##%#%J..r)c^\rSrSr%Sr0\R E\\/\ "\ SSSS9\ "S15/S.Er\\ S 'SS S S S SSSS\ RSSSSS.U4SjjrSrSrU=r$)ria4Randomized search on hyper parameters. The search strategy starts evaluating all the candidates with a small amount of resources and iteratively selects the best candidates, using more and more resources. The candidates are sampled at random from the parameter space and the number of sampled candidates is determined by ``n_candidates``. Read more in the :ref:`User guide`. .. note:: This estimator is still **experimental** for now: the predictions and the API might change without any deprecation cycle. To use it, you need to explicitly import ``enable_halving_search_cv``:: >>> # explicitly require this experimental feature >>> from sklearn.experimental import enable_halving_search_cv # noqa >>> # now you can import normally from model_selection >>> from sklearn.model_selection import HalvingRandomSearchCV Parameters ---------- estimator : estimator object This is assumed to implement the scikit-learn estimator interface. Either estimator needs to provide a ``score`` function, or ``scoring`` must be passed. param_distributions : dict or list of dicts Dictionary with parameters names (`str`) as keys and distributions or lists of parameters to try. Distributions must provide a ``rvs`` method for sampling (such as those from scipy.stats.distributions). If a list is given, it is sampled uniformly. If a list of dicts is given, first a dict is sampled uniformly, and then a parameter is sampled using that dict as above. n_candidates : "exhaust" or int, default="exhaust" The number of candidate parameters to sample, at the first iteration. Using 'exhaust' will sample enough candidates so that the last iteration uses as many resources as possible, based on `min_resources`, `max_resources` and `factor`. In this case, `min_resources` cannot be 'exhaust'. factor : int or float, default=3 The 'halving' parameter, which determines the proportion of candidates that are selected for each subsequent iteration. For example, ``factor=3`` means that only one third of the candidates are selected. resource : ``'n_samples'`` or str, default='n_samples' Defines the resource that increases with each iteration. By default, the resource is the number of samples. It can also be set to any parameter of the base estimator that accepts positive integer values, e.g. 'n_iterations' or 'n_estimators' for a gradient boosting estimator. In this case ``max_resources`` cannot be 'auto' and must be set explicitly. max_resources : int, default='auto' The maximum number of resources that any candidate is allowed to use for a given iteration. By default, this is set ``n_samples`` when ``resource='n_samples'`` (default), else an error is raised. min_resources : {'exhaust', 'smallest'} or int, default='smallest' The minimum amount of resource that any candidate is allowed to use for a given iteration. Equivalently, this defines the amount of resources `r0` that are allocated for each candidate at the first iteration. - 'smallest' is a heuristic that sets `r0` to a small value: - ``n_splits * 2`` when ``resource='n_samples'`` for a regression problem - ``n_classes * n_splits * 2`` when ``resource='n_samples'`` for a classification problem - ``1`` when ``resource != 'n_samples'`` - 'exhaust' will set `r0` such that the **last** iteration uses as much resources as possible. Namely, the last iteration will use the highest value smaller than ``max_resources`` that is a multiple of both ``min_resources`` and ``factor``. In general, using 'exhaust' leads to a more accurate estimator, but is slightly more time consuming. 'exhaust' isn't available when `n_candidates='exhaust'`. Note that the amount of resources used at each iteration is always a multiple of ``min_resources``. aggressive_elimination : bool, default=False This is only relevant in cases where there isn't enough resources to reduce the remaining candidates to at most `factor` after the last iteration. If ``True``, then the search process will 'replay' the first iteration for as long as needed until the number of candidates is small enough. This is ``False`` by default, which means that the last iteration may evaluate more than ``factor`` candidates. See :ref:`aggressive_elimination` for more details. cv : int, cross-validation generator or an iterable, default=5 Determines the cross-validation splitting strategy. Possible inputs for cv are: - integer, to specify the number of folds in a `(Stratified)KFold`, - :term:`CV splitter`, - An iterable yielding (train, test) splits as arrays of indices. For integer/None inputs, if the estimator is a classifier and ``y`` is either binary or multiclass, :class:`StratifiedKFold` is used. In all other cases, :class:`KFold` is used. These splitters are instantiated with `shuffle=False` so the splits will be the same across calls. Refer :ref:`User Guide ` for the various cross-validation strategies that can be used here. .. note:: Due to implementation details, the folds produced by `cv` must be the same across multiple calls to `cv.split()`. For built-in `scikit-learn` iterators, this can be achieved by deactivating shuffling (`shuffle=False`), or by setting the `cv`'s `random_state` parameter to an integer. scoring : str or callable, default=None Scoring method to use to evaluate the predictions on the test set. - str: see :ref:`scoring_string_names` for options. - callable: a scorer callable object (e.g., function) with signature ``scorer(estimator, X, y)``. See :ref:`scoring_callable` for details. - `None`: the `estimator`'s :ref:`default evaluation criterion ` is used. refit : bool or callable, default=True Refit an estimator using the best found parameters on the whole dataset. Where there are considerations other than maximum score in choosing a best estimator, ``refit`` can be set to a function which returns the selected ``best_index_`` given ``cv_results_``. In that case, the ``best_estimator_`` and ``best_params_`` will be set according to the returned ``best_index_`` while the ``best_score_`` attribute will not be available. The refitted estimator is made available at the ``best_estimator_`` attribute and permits using ``predict`` directly on this ``HalvingRandomSearchCV`` instance. See :ref:`this example ` for an example of how to use ``refit=callable`` to balance model complexity and cross-validated score. error_score : 'raise' or numeric Value to assign to the score if an error occurs in estimator fitting. If set to 'raise', the error is raised. If a numeric value is given, FitFailedWarning is raised. This parameter does not affect the refit step, which will always raise the error. Default is ``np.nan``. return_train_score : bool, default=False If ``False``, the ``cv_results_`` attribute will not include training scores. Computing training scores is used to get insights on how different parameter settings impact the overfitting/underfitting trade-off. However computing the scores on the training set can be computationally expensive and is not strictly required to select the parameters that yield the best generalization performance. random_state : int, RandomState instance or None, default=None Pseudo random number generator state used for subsampling the dataset when `resources != 'n_samples'`. Also used for random uniform sampling from lists of possible values instead of scipy.stats distributions. Pass an int for reproducible output across multiple function calls. See :term:`Glossary `. n_jobs : int or None, default=None Number of jobs to run in parallel. ``None`` means 1 unless in a :obj:`joblib.parallel_backend` context. ``-1`` means using all processors. See :term:`Glossary ` for more details. verbose : int Controls the verbosity: the higher, the more messages. Attributes ---------- n_resources_ : list of int The amount of resources used at each iteration. n_candidates_ : list of int The number of candidate parameters that were evaluated at each iteration. n_remaining_candidates_ : int The number of candidate parameters that are left after the last iteration. It corresponds to `ceil(n_candidates[-1] / factor)` max_resources_ : int The maximum number of resources that any candidate is allowed to use for a given iteration. Note that since the number of resources used at each iteration must be a multiple of ``min_resources_``, the actual number of resources used at the last iteration may be smaller than ``max_resources_``. min_resources_ : int The amount of resources that are allocated for each candidate at the first iteration. n_iterations_ : int The actual number of iterations that were run. This is equal to ``n_required_iterations_`` if ``aggressive_elimination`` is ``True``. Else, this is equal to ``min(n_possible_iterations_, n_required_iterations_)``. n_possible_iterations_ : int The number of iterations that are possible starting with ``min_resources_`` resources and without exceeding ``max_resources_``. n_required_iterations_ : int The number of iterations that are required to end up with less than ``factor`` candidates at the last iteration, starting with ``min_resources_`` resources. This will be smaller than ``n_possible_iterations_`` when there isn't enough resources. cv_results_ : dict of numpy (masked) ndarrays A dict with keys as column headers and values as columns, that can be imported into a pandas ``DataFrame``. It contains lots of information for analysing the results of a search. Please refer to the :ref:`User guide` for details. best_estimator_ : estimator or dict Estimator that was chosen by the search, i.e. estimator which gave highest score (or smallest loss if specified) on the left out data. Not available if ``refit=False``. best_score_ : float Mean cross-validated score of the best_estimator. best_params_ : dict Parameter setting that gave the best results on the hold out data. best_index_ : int The index (of the ``cv_results_`` arrays) which corresponds to the best candidate parameter setting. The dict at ``search.cv_results_['params'][search.best_index_]`` gives the parameter setting for the best model, that gives the highest mean score (``search.best_score_``). scorer_ : function or a dict Scorer function used on the held out data to choose the best parameters for the model. n_splits_ : int The number of cross-validation splits (folds/iterations). refit_time_ : float Seconds used for refitting the best model on the whole dataset. This is present only if ``refit`` is not False. multimetric_ : bool Whether or not the scorers compute several metrics. classes_ : ndarray of shape (n_classes,) The classes labels. This is present only if ``refit`` is specified and the underlying estimator is a classifier. n_features_in_ : int Number of features seen during :term:`fit`. Only defined if `best_estimator_` is defined (see the documentation for the `refit` parameter for more details) and that `best_estimator_` exposes `n_features_in_` when fit. .. versionadded:: 0.24 feature_names_in_ : ndarray of shape (`n_features_in_`,) Names of features seen during :term:`fit`. Only defined if `best_estimator_` is defined (see the documentation for the `refit` parameter for more details) and that `best_estimator_` exposes `feature_names_in_` when fit. .. versionadded:: 1.0 See Also -------- :class:`HalvingGridSearchCV`: Search over a grid of parameters using successive halving. Notes ----- The parameters selected are those that maximize the score of the held-out data, according to the scoring parameter. All parameter combinations scored with a NaN will share the lowest rank. Examples -------- >>> from sklearn.datasets import load_iris >>> from sklearn.ensemble import RandomForestClassifier >>> from sklearn.experimental import enable_halving_search_cv # noqa >>> from sklearn.model_selection import HalvingRandomSearchCV >>> from scipy.stats import randint >>> import numpy as np ... >>> X, y = load_iris(return_X_y=True) >>> clf = RandomForestClassifier(random_state=0) >>> np.random.seed(0) ... >>> param_distributions = {"max_depth": [3, None], ... "min_samples_split": randint(2, 11)} >>> search = HalvingRandomSearchCV(clf, param_distributions, ... resource='n_estimators', ... max_resources=10, ... random_state=0).fit(X, y) >>> search.best_params_ # doctest: +SKIP {'max_depth': None, 'min_samples_split': 10, 'n_estimators': 9} rNrYrZr])param_distributionsr|rfrir,r\r^FrhT)r|rdrcrarbrerlr`rkrnror$rjrmcR>[TU]UU UU UU UU U UUUUUS9 X lX0lgr)rqr'rr|)r%rrrr|rdrcrarbrerlr`rkrnror$rjrmrss r&r'HalvingRandomSearchCV.__init__sO(  %#1''#9  $7 (r)cURnUS:XaURUR-n[URUUR S9$)Nr])r$)r|rr}rrr$)r%n_candidates_first_iters r&r0HalvingRandomSearchCV._generate_candidate_params=sQ"&"3"3 "i /'+&9&9T=P=P&P #  $ $ #**  r))r|r)r6r7r8r9r:rWrfrrrrrrr?rr'rr;rrs@r&rrszx $  6 6$ $d| Xq$y 9  { # $D $ FF%%)%)N   r))(abcrmathrrrnumbersrrnumpyr?baser r metrics._scorerr utilsr utils._param_validationrrutils.multiclassrutils.validationrrrr_searchr_splitrr__all__rrUrWrrr<r)r&rsy!!"..:;:-!5 "9 :&&6 ?r Lr j W./W.t w 1w r)