def test_gaussian_noise():
# float
X_new, y_new = resreg.gaussian_noise(X, y, relevance, relevance_threshold=0.5,
delta=0.1, over=0.5, under=0.5, random_state=0)
assert round(np.sum(y_new), 3) == 0.450
# balance
X_new, y_new = resreg.gaussian_noise(X, y, relevance, relevance_threshold=0.5,
delta=0.1, over='balance', random_state=0)
assert round(np.sum(y_new), 3) == 2.504
# extreme
X_new, y_new = resreg.gaussian_noise(X, y, relevance, relevance_threshold=0.5,
delta=0.1, over='extreme', random_state=0)
assert round(np.sum(y_new), 3) == 4.541
# average
X_new, y_new = resreg.gaussian_noise(X, y, relevance, relevance_threshold=0.5,
delta=0.1, over='average', random_state=0)
assert round(np.sum(y_new), 3) == 5.669
# errors
with pytest.raises(Exception):
X_new, y_new = resreg.gaussian_noise(X, y, relevance, relevance_threshold=0.5,
delta=0.1, over=0.5, under=1)
with pytest.raises(Exception):
X_new, y_new = resreg.gaussian_noise(X, y, relevance, relevance_threshold=0.5,
delta=0.1, over=0.5, under=1.5)
with pytest.raises(Exception):
X_new, y_new = resreg.gaussian_noise(X, y, relevance, relevance_threshold=0.5,
delta=0.1, over=0, under=1)
X_train, y_train = resreg.smoter(X_train,
y_train,
relevance,
relevance_threshold=0.5,
k=k,
over=sample_method,
random_state=rrr)
reg.fit(X_train, y_train)
elif strategy == 'GN':
cl, ch, sample_method, delta = param
relevance = resreg.sigmoid_relevance(y_train, cl=cl, ch=ch)
X_train, y_train = resreg.gaussian_noise(X_train,
y_train,
relevance,
relevance_threshold=0.5,
delta=delta,
over=sample_method,
random_state=rrr)
reg.fit(X_train, y_train)
elif strategy == 'WERCS':
cl, ch, over, under = param
relevance = resreg.sigmoid_relevance(y_train, cl=cl, ch=ch)
X_train, y_train = resreg.wercs(X_train,
y_train,
relevance,
over=over,
under=under,
noise=False,
random_state=rrr)
k=10,
over='average',
random_state=0)
sns.kdeplot(y_train,
bw=7,
linewidth=lw,
label=strategy,
color='red',
linestyle=style2)
elif strategy == 'GN':
relevance = resreg.sigmoid_relevance(y_train, cl=None, ch=72.2)
X_train, y_train = resreg.gaussian_noise(X_train,
y_train,
relevance=relevance,
relevance_threshold=0.5,
delta=0.5,
over='balance',
random_state=0)
sns.kdeplot(y_train,
bw=7,
linewidth=lw,
label=strategy,
color='magenta',
linestyle=style2)
elif strategy == 'WERCS':
relevance = resreg.sigmoid_relevance(y_train, cl=None, ch=72.2)
X_train, y_train = resreg.wercs(X_train,
y_train,
relevance=relevance,
