chemicalchecker.tool.targetmate.evaluation.validation.Performances

class Performances(is_classifier)

Bases: object

Calculate performances

Methods

as_dict
classifier_performances	Calculate standard prediction performance metrics.
compute
regressor_performances	TO-DO
truth_inv	Given an active (1) / inactive (0) dataset, assign 1 to inactives

classifier_performances(Yt, Yp)

Calculate standard prediction performance metrics. In addition, it calculates the corresponding weights. For the moment, AUROC, AUPR and BEDROC are used. :param Yt: Truth data (binary). :type Yt: list of lists :param Yp: Prediction scores (probabilities). :type Yp: list of lists

regressor_performances(yt, yp)

TO-DO

static truth_inv(yt)

Given an active (1) / inactive (0) dataset, assign 1 to inactives