RandomizedSearchCV
#include <Skigen/ModelSelection>
template <typename Estimator, typename Scalar = double>
class Skigen::RandomizedSearchCV(estimator, param_distributions, n_iter=10, cv=5, refit=true, random_state=std::nullopt)
Randomised search over a parameter grid with cross-validation.
Mirrors sklearn.model_selection.RandomizedSearchCV.
Samples n_iter random parameter combinations from the supplied grid (with replacement) and evaluates each via K-fold cross-validation. Unlike GridSearchCV, only a fixed budget of combinations is tried.
Attributes:
-
best_estimator : Estimator
-
best_params : ParameterDict
-
best_score : Scalar
-
best_index : std::size_t
-
cv_results_params : const std::vector< ParameterDict > &
-
cv_results_mean_score : const std::vector< Scalar > &