Merge pull request #69 from mohitcek/main

Unit tests for classification metrics

tests/test_classification.py

@@ -0,0 +1,59 @@
+import numpy as np
+
+from langfair.metrics.classification import ClassificationMetrics
+from langfair.metrics.classification.metrics import (
+    FalseDiscoveryRateParity,
+    FalseNegativeRateParity,
+    FalseOmissionRateParity,
+    FalsePositiveRateParity,
+    PredictedPrevalenceRateParity,
+)
+
+groups = np.array([0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1])
+y_pred = np.array([0, 0, 1, 1, 0, 1, 0, 1, 0, 0, 1, 1, 0, 1, 0, 1, 0])
+y_true = np.array([0, 1, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 1, 0, 1, 1])
+
+
+def test_falsediscoveryrateparity():
+    fdrp = FalseDiscoveryRateParity()
+    x1 = fdrp.evaluate(groups=groups, y_pred=y_pred, y_true=y_true)
+    x2 = fdrp.evaluate(groups=groups, y_pred=y_pred, y_true=y_true, ratio=True)
+    assert np.round(x1, 5) == 0.25 and np.round(x2, 5) == 2.0
+
+def test_falsenegativerateparity():
+    fnrp = FalseNegativeRateParity()
+    x1 = fnrp.evaluate(groups=groups, y_pred=y_pred, y_true=y_true)
+    x2 = fnrp.evaluate(groups=groups, y_pred=y_pred, y_true=y_true, ratio=True)
+    assert np.round(x1, 5) == 0.0 and np.round(x2, 5) == 1.0
+
+def test_falseomissionrateparity():
+    forp = FalseOmissionRateParity()
+    x1 = forp.evaluate(groups=groups, y_pred=y_pred, y_true=y_true)
+    x2 = forp.evaluate(groups=groups, y_pred=y_pred, y_true=y_true, ratio=True)
+    assert np.round(x1, 5) == 0.1 and np.round(x2, 5) == 0.83333
+
+def test_falsepositiverateparity():
+    fprp = FalsePositiveRateParity()
+    x1 = fprp.evaluate(groups=groups, y_pred=y_pred, y_true=y_true)
+    x2 = fprp.evaluate(groups=groups, y_pred=y_pred, y_true=y_true, ratio=True)
+    assert np.round(x1, 5) == 0.16667 and np.round(x2, 5) == 1.5
+
+def test_predictedprevalencerateparity():
+    pprp = PredictedPrevalenceRateParity()
+    x1 = pprp.evaluate(groups=groups, y_pred=y_pred)
+    x2 = pprp.evaluate(groups=groups, y_pred=y_pred, ratio=True)
+    assert np.round(x1, 5) == 0.05556 and np.round(x2, 5) == 1.125
+
+def test_classificationmetrics1():
+    cm = ClassificationMetrics(metric_type='all')
+    x = cm.evaluate(groups=groups, y_pred=y_pred, y_true=y_true)
+    assert x == {'FalseNegativeRateParity': 0.0, 'FalseOmissionRateParity': 0.09999999999999998, 
+                 'FalsePositiveRateParity': 0.16666666666666669, 'FalseDiscoveryRateParity': 0.25, 
+                 'PredictedPrevalenceRateParity': 0.05555555555555558}
+
+def test_classificationmetrics2():
+    cm = ClassificationMetrics(metric_type='all')
+    x = cm.evaluate(groups=groups, y_pred=y_pred, y_true=y_true, ratio=True)
+    assert x == {'FalseNegativeRateParity': 1.0, 'FalseOmissionRateParity': 0.8333333333333334, 
+                 'FalsePositiveRateParity': 1.5, 'FalseDiscoveryRateParity': 2.0, 
+                 'PredictedPrevalenceRateParity': 1.125}