factory module¶

The factory to create the machine learning algorithms.

This module contains a factory to instantiate a MLAlgo from its class name. This factory also provides a list of available machine learning algorithms and allows to test if a machine learning algorithm is available.

Classes:

MLAlgoFactory()

This factory instantiates a MLAlgo from its class name.

class gemseo.mlearning.core.factory.MLAlgoFactory[source]¶

Bases: object

This factory instantiates a MLAlgo from its class name.

The class can be either internal to GEMSEO or external. In this second case, it can be either implemented in a module referenced in the “GEMSEO_PATH” or in a module The class can be either internal to GEMSEO or external. In the second case, it can be either implemented in a module referenced in the GEMSEO_PATH environment variable or in a module starting with “gemseo_” and referenced in the PYTHONPATH environment variable.

Methods:

`create`(ml_algo, **options)	Create an instance of a machine learning algorithm.
`is_available`(ml_algo)	Check the availability of a machine learning algorithm.
`load`(directory)	Load an instance of machine learning algorithm from the disk.

Attributes:

models

The available machine learning algorithms.

Return type: None

create(ml_algo, **options)[source]¶

Create an instance of a machine learning algorithm.

Parameters

ml_algo (str) – The name of a machine learning algorithm (its class name).
**options (Optional[Union[gemseo.core.dataset.Dataset, Dict[str, gemseo.mlearning.transform.transformer.Transformer], Any]]) – The options of the machine learning algorithm.

Returns

The instance of the machine learning algorithm.

Return type

gemseo.mlearning.core.ml_algo.MLAlgo

is_available(ml_algo)[source]¶

Check the availability of a machine learning algorithm.

Parameters: ml_algo (str) – The name of a machine learning algorithm (its class name).
Returns: Whether the machine learning algorithm is available.
Return type: bool

load(directory)[source]¶

Load an instance of machine learning algorithm from the disk.

Parameters: directory (Union[str, pathlib.Path]) – The name of the directory containing an instance of a machine learning algorithm.
Returns: The instance of the machine learning algorithm.
Return type: gemseo.mlearning.core.ml_algo.MLAlgo

property models¶: The available machine learning algorithms.