def test_linear_dependence_pd():
X = np.array([
[
0,
0,
1,
1,
],
[
1,
1,
2,
2,
],
[
0.1,
0.2,
1.2,
1.1,
],
]).T
df = pd.DataFrame(X, columns=["a", "b", "c"])
X_tfm = (CorrelationRemover(
sensitive_feature_ids=["a"]).fit(df).transform(df))
assert X_tfm.shape[1] == 2
assert np.allclose(X_tfm[:, 0], 1.5)
def test_linear_dependence():
X = np.array([
[
0,
0,
1,
1,
],
[
1,
1,
2,
2,
],
[
0.1,
0.2,
1.2,
1.1,
],
]).T
X_tfm = CorrelationRemover(sensitive_feature_ids=[0]).fit(X).transform(X)
assert X_tfm.shape[1] == 2
assert np.allclose(X_tfm[:, 0], 1.5)
def test_estimator_checks(test_fn):
test_fn(CorrelationRemover.__name__,
CorrelationRemover(sensitive_feature_ids=[]))
test_fn(CorrelationRemover.__name__,
CorrelationRemover(sensitive_feature_ids=[0]))
# Copyright (c) Microsoft Corporation and Fairlearn contributors.
# Licensed under the MIT License.
import numpy as np
import pandas as pd
from sklearn.utils.estimator_checks import parametrize_with_checks
from fairlearn.preprocessing import CorrelationRemover
@parametrize_with_checks([
CorrelationRemover(sensitive_feature_ids=[]),
CorrelationRemover(sensitive_feature_ids=[0]),
])
def test_sklearn_compatible_estimator(estimator, check):
check(estimator)
def test_linear_dependence():
X = np.array([
[
0,
0,
1,
1,
],
[
1,
1,
2,
2,
"race_Asian", 'race_Caucasian', 'race_Hispanic', 'race_Other',
'race_Unknown', 'had_inpatient_days_False', 'medicare_False'
],
axis=1)
X_raw = X_raw[[
'time_in_hospital', 'had_inpatient_days_True', 'medicare_True',
'race_AfricanAmerican'
]]
# %%
# We are now going to fit the CorrelationRemover to the data,
# and transform it. The transformed array will be placed back
# in a Pandas DataFrame, for plotting purposes.
cr = CorrelationRemover(sensitive_feature_ids=["race_AfricanAmerican"])
X_cr = cr.fit_transform(X_raw)
X_cr = pd.DataFrame(
X_cr,
columns=['time_in_hospital', 'had_inpatient_days_True', 'medicare_True'])
X_cr["race_AfricanAmerican"] = X_raw["race_AfricanAmerican"]
cr_alpha = CorrelationRemover(sensitive_feature_ids=['race_AfricanAmerican'],
alpha=0.5)
X_cr_alpha = cr_alpha.fit_transform(X_raw)
X_cr_alpha = pd.DataFrame(
X_cr_alpha,
columns=['time_in_hospital', 'had_inpatient_days_True', 'medicare_True'])
X_cr_alpha["race_AfricanAmerican"] = X_raw["race_AfricanAmerican"]
# %%