def test_no_predict_proba_before_fit(self):
gs = GridSearch(self.estimator, self.disparity_criterion,
sample_weight_name=self.sample_weight_name)
X, _, _ = _quick_data()
with pytest.raises(NotFittedError) as execInfo:
gs.predict_proba(X)
assert not_fitted_error_msg.format(GridSearch.__name__) == execInfo.value.args[0]
def test_no_predict_proba_before_fit(self):
gs = GridSearch(self.estimator, self.disparity_criterion)
X, _, _ = self._quick_data()
message = str("Must call fit before attempting to make predictions")
with pytest.raises(NotFittedException) as execInfo:
gs.predict_proba(X)
assert message == execInfo.value.args[0]
def test_demographicparity_fair_uneven_populations(A_two_dim):
# Variant of test_demographicparity_already_fair, which has unequal
# populations in the two classes
# Also allow the threshold to be adjustable
score_threshold = 0.625
number_a0 = 4
number_a1 = 4
a0_label = 17
a1_label = 37
X, Y, A = _simple_threshold_data(number_a0, number_a1,
score_threshold, score_threshold,
a0_label, a1_label, A_two_dim)
target = GridSearch(LogisticRegression(solver='liblinear', fit_intercept=True),
constraints=DemographicParity(),
grid_size=11)
target.fit(X, Y, sensitive_features=A)
assert len(target.all_results) == 11
test_X = pd.DataFrame({"actual_feature": [0.2, 0.7],
"sensitive_features": [a0_label, a1_label],
"constant_ones_feature": [1, 1]})
sample_results = target.predict(test_X)
sample_proba = target.predict_proba(test_X)
assert np.allclose(sample_proba, [[0.53748641, 0.46251359], [0.46688736, 0.53311264]])
sample_results = target.all_results[0].predictor.predict(test_X)
assert np.array_equal(sample_results, [1, 0])
def test_demographicparity_fair_uneven_populations_with_grid_offset(
A_two_dim, offset):
# Grid of Lagrangian multipliers has some initial offset
score_threshold = 0.625
number_a0 = 4
number_a1 = 4
a0_label = 17
a1_label = 37
grid_size = 11
iterables = [["+", "-"], ["all"], [a0_label, a1_label]]
midx = pd.MultiIndex.from_product(iterables,
names=["sign", "event", "group_id"])
grid_offset = pd.Series(offset, index=midx)
X, Y, A = _simple_threshold_data(
number_a0,
number_a1,
score_threshold,
score_threshold,
a0_label,
a1_label,
A_two_dim,
)
grid_search = GridSearch(
LogisticRegression(solver="liblinear", fit_intercept=True),
constraints=DemographicParity(),
grid_size=grid_size,
grid_offset=grid_offset,
)
grid_search.fit(X, Y, sensitive_features=A)
assert_n_grid_search_results(grid_size, grid_search)
test_X = pd.DataFrame({
"actual_feature": [0.2, 0.7],
"sensitive_features": [a0_label, a1_label],
"constant_ones_feature": [1, 1],
})
sample_results = grid_search.predict(test_X)
assert np.array_equal(sample_results, [0, 1])
sample_proba = grid_search.predict_proba(test_X)
assert np.allclose(sample_proba,
[[0.55069845, 0.44930155], [0.41546008, 0.58453992]])
sample_results = grid_search.predictors_[0].predict(test_X)
assert np.array_equal(sample_results, [1, 0])
