ForestSearch

class anomalearn.algorithms.tuning.hyperparameter.ForestSearch.ForestSearch(parameter_space: list[skopt.space.space.Categorical | skopt.space.space.Integer | skopt.space.space.Real], saving_folder: str | PathLike, saving_filename: str)

Bases: SkoptSearchABC

Wrapper for the forest_minimize search of skopt.

Inherited attributes

parameter_space

save_filename

save_folder

List of inherited methods

get_results()	Get search results.
search(x, y, objective_function[, ...])	param kwargs: It may have a keyword "skopt_kwargs" specifying the additional named

List of inherited decorated methods

get_results()	Get search results.
search(x, y, objective_function[, ...])	param kwargs: It may have a keyword "skopt_kwargs" specifying the additional named

Inherited methods

get_results() → IHyperparameterSearchResults

Get search results.

Returns:

search_results – The results of the last search or the specified search at initialization.

Return type:

IHyperparameterSearchResults

search(x, y, objective_function: Callable[[ndarray, ndarray, ndarray, ndarray, dict], float], cross_val_generator: ICrossValidation | None = None, train_test_data: bool = False, load_checkpoints: bool = False, *args, **kwargs) → IHyperparameterSearchResults

Parameters:

kwargs – It may have a keyword “skopt_kwargs” specifying the additional named arguments to pass to the skopt optimization function. It cannot have one of “func”, “dimensions”, “x0” or “y0” since they are specified by the wrapper to manage checkpoints and search.

Inherited decorated methods

get_results() → IHyperparameterSearchResults

Get search results.

Returns:

search_results – The results of the last search or the specified search at initialization.

Return type:

IHyperparameterSearchResults

search(x, y, objective_function: Callable[[ndarray, ndarray, ndarray, ndarray, dict], float], cross_val_generator: ICrossValidation | None = None, train_test_data: bool = False, load_checkpoints: bool = False, *args, **kwargs) → IHyperparameterSearchResults

Parameters:

kwargs – It may have a keyword “skopt_kwargs” specifying the additional named arguments to pass to the skopt optimization function. It cannot have one of “func”, “dimensions”, “x0” or “y0” since they are specified by the wrapper to manage checkpoints and search.