sciquence.postprocessing.binarize_classwise

sciquence.postprocessing.binarize_classwise(X, thresholds)

Binarization performed classwise.

Parameters:

• X (numpy.ndarray) – Probabilities vector
• thresholds (list of float or numpy.ndarray) – Binarization thresholds for all the classes

Examples

>>> import numpy as np
>>> X = np.array(
>>> [[ 0.04344385  0.24317802  0.81423947],
>>> [ 0.30503777  0.08385118  0.48402043],
>>> [ 0.38695257  0.64501778  0.19023201],
>>> [ 0.49452506  0.35440145  0.74149338],
>>> [ 0.25147325  0.14294654  0.6648142 ],
>>> [ 0.99852846  0.75026559  0.43106003],
>>> [ 0.33369685  0.41158767  0.86865335],
>>> [ 0.07741532  0.90428353  0.87152301],
>>> [ 0.79609158  0.47617837  0.1890651 ],
>>> [ 0.14287567  0.52800364  0.10957203]]
>>> )
>>> X_binarized = ClasswiseBinarizer(thresholds=[.5, .4, .3]).transform(X)
>>> print X_binarized
>>> [[ 0.  0.  1.],
>>> [ 0.  0.  1.],
>>> [ 0.  1.  0.],
>>> [ 0.  0.  1.],
>>> [ 0.  0.  1.],
>>> [ 1.  1.  1.],
>>> [ 0.  1.  1.],
>>> [ 0.  1.  1.],
>>> [ 1.  1.  0.],
>>> [ 0.  1.  0.]]