def test_numerical_user_splits_fixed():
user_splits = [4, 7, 7.1, 10, 16, 20, 23]
with raises(ValueError):
user_splits_fixed = [True, True, True, True, False, False, False]
optb = ContinuousOptimalBinning(user_splits_fixed=user_splits_fixed)
optb.fit(x, y)
with raises(TypeError):
user_splits_fixed = (False, False, False, False, False, True, False)
optb = ContinuousOptimalBinning(user_splits=user_splits,
user_splits_fixed=user_splits_fixed)
optb.fit(x, y)
with raises(ValueError):
user_splits_fixed = [0, 0, 0, 0, 0, 1, 0]
optb = ContinuousOptimalBinning(user_splits=user_splits,
user_splits_fixed=user_splits_fixed)
optb.fit(x, y)
with raises(ValueError):
user_splits_fixed = [False, False, False, False]
optb = ContinuousOptimalBinning(user_splits=user_splits,
user_splits_fixed=user_splits_fixed)
optb.fit(x, y)
user_splits_fixed = [True, True, True, True, False, False, False]
optb = ContinuousOptimalBinning(user_splits=user_splits,
user_splits_fixed=user_splits_fixed)
optb.fit(x, y)
assert optb.status == "INFEASIBLE"
def test_numerical_default_transform():
optb = ContinuousOptimalBinning()
with raises(NotFittedError):
x_transform = optb.transform(x)
optb.fit(x, y)
x_transform = optb.transform([0.2, 4.1, 7.2, 26])
assert x_transform == approx([39.718, 39.718, 25.56067416, 11.82978723],
rel=1e-6)
def test_numerical_default():
optb = ContinuousOptimalBinning()
optb.fit(x, y)
assert optb.status == "OPTIMAL"
assert optb.splits == approx([
4.6500001, 5.49499989, 6.86500001, 9.7249999, 11.67499971, 13.0999999,
16.08500004, 19.89999962, 23.31500053
],
rel=1e-6)
def test_numerical_default():
optb = ContinuousOptimalBinning()
optb.fit(x, y)
assert optb.status == "OPTIMAL"
assert optb.splits == approx([4.6500001, 5.49499989, 6.86500001, 9.7249999,
11.67499971, 13.0999999, 16.08500004,
19.89999962, 23.31500053],
rel=1e-6)
optb.binning_table.build()
optb.binning_table.analysis()
optb.binning_table.plot(savefig="test_continuous_binning.png")
optb.binning_table.plot(add_special=False,
savefig="test_continuous_binning_no_special.png")
optb.binning_table.plot(add_missing=False,
savefig="test_continuous_binning_no_missing.png")
def test_default_transform_continuous():
data = load_boston()
variable_names = data.feature_names
X = data.data
y = data.target
process = BinningProcess(variable_names)
process.fit(X, y)
X_transform = process.transform(X)
optb = process.get_binned_variable(variable_names[0])
assert isinstance(optb, ContinuousOptimalBinning)
optb = ContinuousOptimalBinning()
x = X[:, 5]
optb.fit(x, y)
assert optb.transform(x) == approx(X_transform[:, 5], rel=1e-6)
def test_params():
with raises(TypeError):
optb = ContinuousOptimalBinning(name=1)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(dtype="nominal")
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(prebinning_method="new_method")
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(max_n_prebins=-2)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(min_prebin_size=0.6)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(min_n_bins=-2)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(max_n_bins=-2.2)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(min_n_bins=3, max_n_bins=2)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(min_bin_size=0.6)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(max_bin_size=-0.6)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(min_bin_size=0.5, max_bin_size=0.3)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(monotonic_trend="new_trend")
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(min_mean_diff=-1.1)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(max_pvalue=1.1)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(max_pvalue_policy="new_policy")
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(cat_cutoff=-0.2)
optb.fit(x, y)
with raises(TypeError):
optb = ContinuousOptimalBinning(user_splits={"a": [1, 2]})
optb.fit(x, y)
with raises(TypeError):
optb = ContinuousOptimalBinning(special_codes={1, 2, 3})
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(split_digits=9)
optb.fit(x, y)
with raises(ValueError):
optb = ContinuousOptimalBinning(time_limit=-2)
optb.fit(x, y)
with raises(TypeError):
optb = ContinuousOptimalBinning(verbose=1)
optb.fit(x, y)
def test_verbose():
optb = ContinuousOptimalBinning(verbose=True)
optb.fit(x, y)
assert optb.status == "OPTIMAL"
def test_numerical_user_splits_non_unique():
user_splits = [4, 7, 7, 10, 16, 20, 23]
optb = ContinuousOptimalBinning(user_splits=user_splits)
with raises(ValueError):
optb.fit(x, y)