def test_pipeline(self):
X, y = samples_generator.make_classification(
n_informative=5, n_redundant=0, random_state=42)
anova_filter = SelectKBest(f_regression, k=5)
clf = SVC(kernel='linear')
anova_svc = Pipeline([('anova', anova_filter),
('svc', clf)])
anova_svc.fit(X, y)
score = anova_svc.score(X,y)
fname = tempfile.NamedTemporaryFile().name
f = EstimatorStore(fname, "w")
path = "anova_svc"
f.store_fit(anova_svc, path, set_params=True)
f.close()
# redefine
anova_filter2 = SelectKBest(f_regression)
clf2 = SVC()
anova_svc2 = Pipeline([('anova', anova_filter2),
('svc', clf2)])
f = EstimatorStore(fname)
f.restore_fit(anova_svc2, path, set_params=True)
score2 = anova_svc2.score(X,y)
assert score == score2
for attr in f.get_fitted_attrs(anova_filter):
assert_array_equal( getattr(anova_filter, attr),
getattr(anova_filter2, attr) )
for attr in f.get_fitted_attrs(clf):
assert_array_equal( getattr(clf, attr),
getattr(clf2, attr) )
def test_parameters(self):
iris = load_iris()
X, y = iris.data, iris.target
estimator = MultinomialNB(alpha=2.0, fit_prior=False)
estimator.fit(X, y)
fname = tempfile.NamedTemporaryFile().name
f = EstimatorStore(fname, "w")
path = "nb"
f.store_fit(estimator, path, set_params=True)
f.close()
estimator = MultinomialNB()
f = EstimatorStore(fname)
f.restore_fit(estimator, path, set_params=True)
assert estimator.alpha == 2.0
assert estimator.fit_prior is False
def test_single_estimator(self):
iris = load_iris()
X, y = iris.data, iris.target
estimator = MultinomialNB()
estimator.fit(X, y)
class_priors = estimator.class_log_prior_
feat_prob = estimator.feature_log_prob_
fname = tempfile.NamedTemporaryFile().name
f = EstimatorStore(fname, "w")
path = "nb"
f.store_fit(estimator, path)
f.close()
estimator = MultinomialNB()
f = EstimatorStore(fname)
f.restore_fit(estimator, path)
assert_array_equal(class_priors, estimator.class_log_prior_ )
assert_array_equal(feat_prob, estimator.feature_log_prob_)