test_volume_calculation.py

import numpy as np
from cwc.evaluation.abstaingaincurve import AbstainGainCurve
from cwc.evaluation.rgp import RGP
from cwc.evaluation.rgrpg import RGRPG


if __name__ == "__main__":
    predicted_labels_python = np.genfromtxt("datasets/predicted_labels_python.csv", delimiter=",", skip_header=True)
    step1_labels = predicted_labels_python[:, 4]
    step1_scores = predicted_labels_python[:, 0]
    step1_reject_scores = step1_scores[step1_labels == 0]
    step1_training_scores = step1_scores[step1_labels == 1]
    training_labels = 1 * (predicted_labels_python[:, 5][step1_labels == 1] == 1)
    step2_training_scores = predicted_labels_python[:, 2][step1_labels == 1]
    # rgp = RGP(step1_reject_scores, step1_training_scores, step2_training_scores, training_labels)
    # rgp.plot()
    # print "Gaintp area: " + str(rgp.calculate_area())
    rgrpg = RGRPG(step1_reject_scores, step1_training_scores, step2_training_scores, training_labels)
    rgrpg.plot_simple_3d()
    # print "RGRPG volume: " + str(rgrpg.calculate_volume())
    # rgrpg.plot_rgrpg_2d()
    # rgrpg.plot_rgrpg_3d(n_recalls=50, n_points_roc=50)

    # ag = AbstainGainCurve(step1_reject_scores, step1_training_scores, step2_training_scores, training_labels)
    # ag.plot()
    # print "Abstain-gain area: " + str(ag.calculate_area())