def testToTFIDFNotSmooth(self):
term_freq = tf.SparseTensor(
[[0, 0], [0, 1], [0, 2], [1, 0], [1, 3]],
[(3/5), (1/5), (1/5), (1/2), (1/2)],
[2, 4])
reduced_term_freq = tf.constant([[2, 1, 1, 1]])
output_tensor = mappers._to_tfidf(term_freq, reduced_term_freq, 2, False)
log_2_over_1 = 1.6931471
self.assertSparseOutput(
expected_indices=[[0, 0], [0, 1], [0, 2], [1, 0], [1, 3]],
expected_values=[(3/5), (1/5)*log_2_over_1, (1/5)*log_2_over_1,
(1/2), (1/2)*log_2_over_1],
expected_shape=[2, 4],
actual_sparse_tensor=output_tensor,
close_values=True)
def testToTFIDF(self):
term_freq = tf.SparseTensor(
[[0, 0], [0, 1], [0, 2], [1, 0], [1, 3]],
[(3/5), (1/5), (1/5), (1/2), (1/2)],
[2, 4])
reduced_term_freq = tf.constant([[2, 1, 1, 1]])
output_tensor = mappers._to_tfidf(term_freq, reduced_term_freq, 2, True)
log_3_over_2 = 0.4054651
self.assertSparseOutput(
expected_indices=[[0, 0], [0, 1], [0, 2], [1, 0], [1, 3]],
expected_values=[0, (1/5)*log_3_over_2, (1/5)*log_3_over_2,
0, (1/2)*log_3_over_2],
expected_shape=[2, 4],
actual_sparse_tensor=output_tensor,
close_values=True)