def test_soft_metrics_computation(self) -> None:
recall_at_precision_dict_l1 = {
0.9: 0.0,
0.8: 0.0,
0.6: 1.0,
0.4: 1.0,
0.2: 1.0
}
recall_at_precision_dict_l2 = {
0.9: 0.5,
0.8: 0.5,
0.6: 0.5,
0.4: 1.0,
0.2: 1.0
}
self.assertMetricsAlmostEqual(
compute_soft_metrics(PREDICTIONS2, LABEL_NAMES2),
{
"label1":
SoftClassificationMetrics(
average_precision=8.0 / 15,
recall_at_precision=recall_at_precision_dict_l1,
),
"label2":
SoftClassificationMetrics(
average_precision=0.7,
recall_at_precision=recall_at_precision_dict_l2,
),
},
)
def test_soft_metrics_computation(self) -> None:
recall_at_precision_dict_l1 = {
0.9: 0.0,
0.8: 0.0,
0.6: 1.0,
0.4: 1.0,
0.2: 1.0
}
decision_thresh_at_precision_dict_l1 = {
0.9: 0.0,
0.8: 0.0,
0.6: -0.7,
0.4: -0.7,
0.2: -0.7,
}
recall_at_precision_dict_l2 = {
0.9: 0.5,
0.8: 0.5,
0.6: 0.5,
0.4: 1.0,
0.2: 1.0
}
decision_thresh_at_precision_dict_l2 = {
0.9: -0.2,
0.8: -0.2,
0.6: -0.2,
0.4: -0.8,
0.2: -0.8,
}
precision_at_recall_dict_l1 = {
0.9: 0.6,
0.8: 0.6,
0.6: 0.6,
0.4: 0.6,
0.2: 0.6
}
decision_thresh_at_recall_dict_l1 = {
0.9: -0.7,
0.8: -0.7,
0.6: -0.7,
0.4: -0.7,
0.2: -0.7,
}
precision_at_recall_dict_l2 = {
0.9: 0.4,
0.8: 0.4,
0.6: 0.4,
0.4: 1.0,
0.2: 1.0
}
decision_thresh_at_recall_dict_l2 = {
0.9: -0.8,
0.8: -0.8,
0.6: -0.8,
0.4: -0.2,
0.2: -0.2,
}
self.assertMetricsAlmostEqual(
compute_soft_metrics(PREDICTIONS2, LABEL_NAMES2),
{
"label1":
SoftClassificationMetrics(
average_precision=8.0 / 15,
recall_at_precision=recall_at_precision_dict_l1,
decision_thresh_at_precision=
decision_thresh_at_precision_dict_l1,
precision_at_recall=precision_at_recall_dict_l1,
decision_thresh_at_recall=decision_thresh_at_recall_dict_l1,
roc_auc=1.0 / 6,
),
"label2":
SoftClassificationMetrics(
average_precision=0.7,
recall_at_precision=recall_at_precision_dict_l2,
decision_thresh_at_precision=
decision_thresh_at_precision_dict_l2,
precision_at_recall=precision_at_recall_dict_l2,
decision_thresh_at_recall=decision_thresh_at_recall_dict_l2,
roc_auc=4.0 / 6,
),
},
)