def test_summary_get_support():
data = load_breast_cancer()
variable_names = data.feature_names
X = data.data
y = data.target
process = BinningProcess(variable_names, min_iv=0.1, max_iv=0.6)
with raises(ValueError):
process.summary()
with raises(ValueError):
process.get_support()
process.fit(X, y, check_input=True)
assert isinstance(process.summary(), pd.DataFrame)
with raises(ValueError):
process.get_support(indices=True, names=True)
assert all(process.get_support() == [
False, False, False, False, False, False, False, False, False, True,
False, True, False, False, True, False, False, False, True, True,
False, False, False, False, False, False, False, False, False, True
])
assert process.get_support(indices=True) == approx([9, 11, 14, 18, 19, 29])
assert all(
process.get_support(names=True) == [
'mean fractal dimension', 'texture error', 'smoothness error',
'symmetry error', 'fractal dimension error',
'worst fractal dimension'
])
def test_summary_get_support():
data = load_breast_cancer()
variable_names = data.feature_names
X = data.data
y = data.target
selection_criteria = {"iv": {"min": 0.1, "max": 0.6,
"strategy": "highest", "top": 10}}
process = BinningProcess(variable_names=variable_names,
selection_criteria=selection_criteria)
with raises(ValueError):
process.summary()
with raises(ValueError):
process.get_support()
process.fit(X, y, check_input=True)
assert isinstance(process.summary(), pd.DataFrame)
with raises(ValueError):
process.get_support(indices=True, names=True)
assert all(process.get_support() == [
False, False, False, False, False, False, False, False, False, True,
False, True, False, False, True, False, False, False, True, True,
False, False, False, False, False, False, False, False, False, True])
assert process.get_support(indices=True) == approx([9, 11, 14, 18, 19, 29])
assert all(process.get_support(names=True) == [
'mean fractal dimension', 'texture error', 'smoothness error',
'symmetry error', 'fractal dimension error',
'worst fractal dimension'])
def test_categorical_variables():
data = load_boston()
variable_names = data.feature_names
X = data.data
y = data.target
process = BinningProcess(variable_names, categorical_variables=["CHAS"])
process.fit(X, y, check_input=True)
df_summary = process.summary()
assert df_summary[df_summary.name ==
"CHAS"]["dtype"].values[0] == "categorical"
def test_auto_modes():
df = pd.DataFrame(data.data, columns=data.feature_names)
binning_fit_params0 = {
v: {
"monotonic_trend": "auto",
"solver": "mip"
}
for v in data.feature_names
}
binning_fit_params1 = {
v: {
"monotonic_trend": "auto_heuristic",
"solver": "mip"
}
for v in data.feature_names
}
binning_fit_params2 = {
v: {
"monotonic_trend": "auto",
"solver": "cp"
}
for v in data.feature_names
}
binning_fit_params3 = {
v: {
"monotonic_trend": "auto_heuristic",
"solver": "cp"
}
for v in data.feature_names
}
process0 = BinningProcess(variable_names,
binning_fit_params=binning_fit_params0)
process1 = BinningProcess(variable_names,
binning_fit_params=binning_fit_params1)
process2 = BinningProcess(variable_names,
binning_fit_params=binning_fit_params2)
process3 = BinningProcess(variable_names,
binning_fit_params=binning_fit_params3)
process0.fit(df, y)
process1.fit(df, y)
process2.fit(df, y)
process3.fit(df, y)
assert process0.summary().iv.sum() == process1.summary().iv.sum()
assert process2.summary().iv.sum() == process3.summary().iv.sum()
assert process0.summary().iv.sum() == process2.summary().iv.sum()