class dask_ml.model_selection.RandomizedSearchCV(estimator, param_distributions, n_iter=10, random_state=None, scoring=None, iid=True, refit=True, cv=None, error_score='raise', return_train_score=False, scheduler=None, n_jobs=- 1, cache_cv=True)

Randomized search on hyper parameters.

RandomizedSearchCV implements a “fit” and a “score” method. It also implements “predict”, “predict_proba”, “decision_function”, “transform” and “inverse_transform” if they are implemented in the estimator used.

The parameters of the estimator used to apply these methods are optimized by cross-validated search over parameter settings.

In contrast to GridSearchCV, not all parameter values are tried out, but rather a fixed number of parameter settings is sampled from the specified distributions. The number of parameter settings that are tried is given by n_iter.

If all parameters are presented as a list, sampling without replacement is performed. If at least one parameter is given as a distribution, sampling with replacement is used. It is highly recommended to use continuous distributions for continuous parameters.

estimatorestimator object.

A object of this type is instantiated for each parameter. This is assumed to implement the scikit-learn estimator interface. Either estimator needs to provide a score function, or scoring must be passed. If a list of dicts is given, first a dict is sampled uniformly, and then a parameter is sampled using that dict as above.


Dictionary with parameters names (string) 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.

n_iterint, default=10

Number of parameter settings that are sampled. n_iter trades off runtime vs quality of the solution.

random_stateint or RandomState

Pseudo random number generator state 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.

scoringstring, callable, list/tuple, dict or None, default: None

A single string or a callable to evaluate the predictions on the test set.

For evaluating multiple metrics, either give a list of (unique) strings or a dict with names as keys and callables as values.

NOTE that when using custom scorers, each scorer should return a single value. Metric functions returning a list/array of values can be wrapped into multiple scorers that return one value each.

If None, the estimator’s default scorer (if available) is used.

iidboolean, default=True

If True, the data is assumed to be identically distributed across the folds, and the loss minimized is the total loss per sample, and not the mean loss across the folds.

cvint, cross-validation generator or an iterable, optional

Determines the cross-validation splitting strategy. Possible inputs for cv are:

  • None, to use the default 3-fold cross validation,

  • integer, to specify the number of folds in a (Stratified)KFold,

  • An object to be used as a cross-validation generator.

  • 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, StratifiedKFold is used. In all other cases, KFold is used.

refitboolean, or string, default=True

Refit an estimator using the best found parameters on the whole dataset.

For multiple metric evaluation, this needs to be a string denoting the scorer is used to find the best parameters for refitting the estimator at the end.

The refitted estimator is made available at the best_estimator_ attribute and permits using predict directly on this GridSearchCV instance.

Also for multiple metric evaluation, the attributes best_index_, best_score_ and best_parameters_ will only be available if refit is set and all of them will be determined w.r.t this specific scorer.

See scoring parameter to know more about multiple metric evaluation.

error_score‘raise’ (default) 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.

return_train_scoreboolean, default=True

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.

Note that for scikit-learn >= 0.19.1, the default of True is deprecated, and a warning will be raised when accessing train score results without explicitly asking for train scores.

schedulerstring, callable, Client, or None, default=None

The dask scheduler to use. Default is to use the global scheduler if set, and fallback to the threaded scheduler otherwise. To use a different scheduler either specify it by name (either “threading”, “multiprocessing”, or “synchronous”), pass in a dask.distributed.Client, or provide a scheduler get function.

n_jobsint, default=-1

Number of jobs to run in parallel. Ignored for the synchronous and distributed schedulers. If n_jobs == -1 [default] all cpus are used. For n_jobs < -1, (n_cpus + 1 + n_jobs) are used.

cache_cvbool, default=True

Whether to extract each train/test subset at most once in each worker process, or every time that subset is needed. Caching the splits can speedup computation at the cost of increased memory usage per worker process.

If True, worst case memory usage is (n_splits + 1) * (X.nbytes + y.nbytes) per worker. If False, worst case memory usage is (n_threads_per_worker + 1) * (X.nbytes + y.nbytes) per worker.

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.

For instance the below given table






















will be represented by a cv_results_ dict of:

'param_kernel': masked_array(data = ['poly', 'poly', 'rbf', 'rbf'],
                                mask = [False False False False]...)
'param_gamma': masked_array(data = [-- -- 0.1 0.2],
                            mask = [ True  True False False]...),
'param_degree': masked_array(data = [2.0 3.0 -- --],
                                mask = [False False  True  True]...),
'split0_test_score'  : [0.8, 0.7, 0.8, 0.9],
'split1_test_score'  : [0.82, 0.5, 0.7, 0.78],
'mean_test_score'    : [0.81, 0.60, 0.75, 0.82],
'std_test_score'     : [0.02, 0.01, 0.03, 0.03],
'rank_test_score'    : [2, 4, 3, 1],
'split0_train_score' : [0.8, 0.7, 0.8, 0.9],
'split1_train_score' : [0.82, 0.7, 0.82, 0.5],
'mean_train_score'   : [0.81, 0.7, 0.81, 0.7],
'std_train_score'    : [0.03, 0.04, 0.03, 0.03],
'mean_fit_time'      : [0.73, 0.63, 0.43, 0.49],
'std_fit_time'       : [0.01, 0.02, 0.01, 0.01],
'mean_score_time'    : [0.007, 0.06, 0.04, 0.04],
'std_score_time'     : [0.001, 0.002, 0.003, 0.005],
'params'             : [{'kernel': 'poly', 'degree': 2}, ...],

NOTE that the key 'params' is used to store a list of parameter settings dict for all the parameter candidates.

The mean_fit_time, std_fit_time, mean_score_time and std_score_time are all in seconds.


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 or dict of floats

Score of best_estimator on the left out data. When using multiple metrics, best_score_ will be a dictionary where the keys are the names of the scorers, and the values are the mean test score for that scorer.


Parameter setting that gave the best results on the hold out data.

best_index_int or dict of ints

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_).

When using multiple metrics, best_index_ will be a dictionary where the keys are the names of the scorers, and the values are the index with the best mean score for that scorer, as described above.

scorer_function or dict of functions

Scorer function used on the held out data to choose the best parameters for the model. A dictionary of {scorer_name: scorer} when multiple metrics are used.


The number of cross-validation splits (folds/iterations).


The parameters selected are those that maximize the score of the left out data, unless an explicit score is passed in which case it is used instead.


>>> import dask_ml.model_selection as dcv
>>> from scipy.stats import expon
>>> from sklearn import svm, datasets
>>> iris = datasets.load_iris()
>>> parameters = {'C': expon(scale=100), 'kernel': ['linear', 'rbf']}
>>> svc = svm.SVC()
>>> clf = dcv.RandomizedSearchCV(svc, parameters, n_iter=100)
>>> clf.fit(iris.data, iris.target)  
RandomizedSearchCV(cache_cv=..., cv=..., error_score=...,
        estimator=SVC(C=..., cache_size=..., class_weight=..., coef0=...,
                      decision_function_shape=..., degree=..., gamma=...,
                      kernel=..., max_iter=..., probability=...,
                      random_state=..., shrinking=..., tol=...,
        iid=..., n_iter=..., n_jobs=..., param_distributions=...,
        random_state=..., refit=..., return_train_score=...,
        scheduler=..., scoring=...)
>>> sorted(clf.cv_results_.keys())  
['mean_fit_time', 'mean_score_time', 'mean_test_score',...
 'mean_train_score', 'param_C', 'param_kernel', 'params',...
 'rank_test_score', 'split0_test_score',...
 'split0_train_score', 'split1_test_score', 'split1_train_score',...
 'split2_test_score', 'split2_train_score',...
 'std_fit_time', 'std_score_time', 'std_test_score', 'std_train_score'...]



Call decision_function on the estimator with the best found parameters.

fit(X[, y, groups])

Run fit with all sets of parameters.


Get metadata routing of this object.


Get parameters for this estimator.


Call inverse_transform on the estimator with the best found params.


Call predict on the estimator with the best found parameters.


Call predict_log_proba on the estimator with the best found parameters.


Call predict_proba on the estimator with the best found parameters.

score(X[, y])

Return the score on the given data, if the estimator has been refit.

set_fit_request(*[, groups])

Request metadata passed to the fit method.


Set the parameters of this estimator.


Call transform on the estimator with the best found parameters.

visualize([filename, format])

Render the task graph for this parameter search using graphviz.

__init__(estimator, param_distributions, n_iter=10, random_state=None, scoring=None, iid=True, refit=True, cv=None, error_score='raise', return_train_score=False, scheduler=None, n_jobs=- 1, cache_cv=True)