Pipelines#

Functions#

`full_pipeline`(dataset_params, model_params, ...)	A pipeline with data retrieval, processing, model training and model evaluation.
`data_exploration_pipeline`()

Full Pipeline#

class mlcompare.full_pipeline(dataset_params, model_params, task_type, save_models='none', save_original=True, save_processed=True, save_directory=None)[source]#

Bases:

A pipeline with data retrieval, processing, model training and model evaluation.

Return type:

None

Parameters:

dataset_params (list[dict[str, Any]] | str | Path)
model_params (list[dict[str, Any]] | str | Path)
task_type (Literal['classification', 'regression'])
save_models (Literal['all', 'best', 'none'])
save_original (bool)
save_processed (bool)
save_directory (str | Path | None)

Args:#

dataset_params (ParamsInput): List containing dataset information. model_params (ParamsInput): List containing model information. task_type (Literal[“classification”, “regression”]): Type of machine learning task to be performed. save_models (Literal[“all”, “best”, “none”], optional): Save all models, only the best model, or no models. Defaults to “none”. save_original (bool, optional): Save original datasets. Defaults to True. save_processed (bool, optional): Save processed datasets. Defaults to True. save_directory (str | Path, optional): Directory to save data, models, and results to. Defaults to “mlc-y-m-DTH-M-S-MS”.

Data Pipeline#

class mlcompare.data_pipeline(dataset_params, save_original=True, save_processed=True, save_directory=None)[source]#

Bases:

A pipeline which only performs data retrieval and/or processing.

Return type:

None

Parameters:

dataset_params (list[dict[str, Any]] | str | Path)
save_original (bool)
save_processed (bool)
save_directory (str | Path | None)

Args:#

dataset_params (ParamsInput): Parameters for loading and processing datasets. save_original (bool, optional): Save original datasets. Defaults to True. save_processed (bool, optional): Save processed datasets. Defaults to True. save_directory (str | Path, optional): Directory to save results to. Defaults to “mlcompare-results-Y-m-dTH-M-S”