quality_measure module¶

Quality measure¶

The quality_measure module implements the concept of quality measures for machine learning algorithms.

This concept is implemented through the MLQualityMeasure class.

class gemseo.mlearning.qual_measure.quality_measure.MLQualityMeasure(algo)[source]¶

Bases: object

Quality measure for machine learning.

Constructor.

evaluate(method='learn', **options)[source]¶

Evaluate quality measure.

Parameters

method (str) – method to estimate the quality measure.
options – options of the estimation method (e.g. ‘test_data’ for the ‘test’ method, ‘n_replicates’ for the boostrap one, …)

Returns

quality measure value.

evaluate_bootstrap(n_replicates=100, multioutput=True)[source]¶

Evaluate quality measure using the bootstrap technique.

Parameters

n_replicates (int) – number of bootstrap replicates. Default: 100.
multioutput (bool) – if True, return the quality measure for each component. Otherwise, average these measures. Default: True.

Returns

quality measure value.

evaluate_kfolds(n_folds=5, multioutput=True)[source]¶

Evaluate quality measure using the k-folds technique.

Parameters

n_folds (int) – number of folds. Default: 5.
multioutput (bool) – if True, return the quality measure for each component. Otherwise, average these measures. Default: True.

Returns

quality measure value.

evaluate_learn(multioutput=True)[source]¶

Evaluate quality measure using the learning dataset.

Parameters: multioutput (bool) – if True, return the quality measure for each output component. Otherwise, average these measures. Default: True.
Returns: quality measure value.

evaluate_loo(multioutput=True)[source]¶

Evaluate quality measure using the leave-one-out technique.

Parameters: multioutput (bool) – if True, return the quality measure for each component. Otherwise, average these measures. Default: True.
Returns: quality measure value.

evaluate_test(test_data, multioutput=True)[source]¶

Evaluate quality measure using a test dataset.

Parameters

test_data (Dataset) – test data.
multioutput (bool) – if True, return the quality measure for each output component. Otherwise, average these measures. Default: True.

Returns

quality measure value.