def testEntailmentClassifierEmptyData(self): # Arrange data, test_data, vectors = testData() vectors.nouns = defaultdict(lambda: {}) neigh = KNeighborsClassifier(n_neighbors=1) classifier = EntailmentClassifier(neigh, vectors) # Act classifier.fit(data) classifier.predict(data)
def testEntailmentClassifier(self): for i in range(3): # Arrange data, test_data, vectors = testData() expected = tuple(x[2] for x in test_data) test_data = [x[:2] for x in test_data] neigh = KNeighborsClassifier(n_neighbors=1) classifier = EntailmentClassifier(neigh, vectors) # Act classifier.fit(data) results = classifier.predict(test_data) # Assert self.assertEqual(tuple(results), expected)
def _make_linsvmDIFF(self): linsvm = LinearSVC() return EntailmentClassifier(linsvm, self.vectors)
def _make_most_frequent(self): dummy = DummyClassifier('most_frequent') return EntailmentClassifier(dummy, self.vectors)
def _make_knn10(self): neigh = KNeighborsClassifier(n_neighbors=10) return EntailmentClassifier(neigh, self.vectors)
def _make_knnP(self): neigh = GridSearchCV(KNeighborsClassifier(), {'n_neighbors': self.params['k']}, score_func=f1_score) return EntailmentClassifier(neigh, self.vectors)