modulation.classification.metrics.lwlrap.one_sample_positive_class_precisions

modulation.classification.metrics.lwlrap.one_sample_positive_class_precisions(scores: array, truths: array) → array

Reference implementation of l$omega$lrap both natively and using sklearn.metrics.

Calculate precisions for each true class for a single sample.

Args:

scores: numpy.array of (num_classes,) giving the individual classifier scores. truths: numpy.array of (num_classes,) bools indicating which classes are true.

Returns:

pos_class_indices: numpy.array of indices of the true classes for this sample. pos_class_precisions: numpy.array of precisions corresponding to each of those

classes.