def test_get_normalised_similarity(self):
minimum = 0.1
maximum = 0.7
beta = 1.0
alpha = 2.0
self.assertEqual(KNN.get_normalised_similarity(minimum, maximum, beta, alpha, 0.2), 7 / 6)
self.assertEqual(KNN.get_normalised_similarity(minimum, maximum, beta, alpha, 0.3), 4 / 3)
self.assertEqual(KNN.get_normalised_similarity(minimum, maximum, beta, alpha, 0.4), 3 / 2)
def get_test_results(self, training, testing):
"""Returns the confusion matrix from the fold validation for the given training and testing sets."""
knn = KNN(training, self.k)
confusion_matrix = ConfusionMatrix(knn.classes)
for document in testing:
predicted = knn.get_prediction(document)
actual = document.c
confusion_matrix.add(actual, predicted)
return confusion_matrix
def test_get_simple_majority_class(self):
documents = [
ClassDocument("A positive document.", "+"),
ClassDocument("Another positive document.", "+"),
ClassDocument("A negative document instead.", "-")
]
self.assertEqual(KNN.get_simple_majority_class(documents), "+")
def test_get_majority_class(self):
documents = [
ClassDocument("pos", "+"),
ClassDocument("pos", "+"),
ClassDocument("neg", "-"),
ClassDocument("neg", "-"),
ClassDocument("neg", "-")
]
similarities = [0.9, 0.8, 0.3, 0.2, 0.1]
self.assertEqual(KNN.get_majority_class(documents, similarities), "+")
def test_get_magnitude(self):
self.assertEqual(KNN.get_magnitude({"hello": 1, "my": 1, "friend": 3}), sqrt(11))
self.assertEqual(KNN.get_magnitude({"hello": 1, "my": 1, "enemy": 2}), sqrt(6))
def test_get_minimum_and_maximum(self):
similarities = [0.14, 0.01, 0.98, 0.43]
self.assertEqual(KNN.get_minimum_and_maximum(similarities), (0.01, 0.98))
def test_get_similarity(self):
document_vector_a = {"hello": 1, "my": 1, "friend": 3}
document_vector_b = {"hello": 1, "my": 1, "enemy": 2}
self.assertEqual(KNN.get_similarity(document_vector_a, document_vector_b), 2 / (sqrt(11) * sqrt(6)))
def test_get_dot_product(self):
document_vector_a = {"hello": 1, "my": 1, "friend": 3}
document_vector_b = {"hello": 1, "my": 1, "enemy": 2}
expected_dot_product = 2
self.assertEqual(KNN.get_dot_product(document_vector_a, document_vector_b), expected_dot_product)