class TFIDFModelTest(unittest.TestCase):
def setUp(self):
self.data = DummyDataset()
self.model = TFIDFModel(verbose = False)
def test_tocsr(self):
# test item/tag matrix conversion
TF_expected = [[1, 1, 1, 1, 1, 1],
[1, 0, 4, 0, 1, 0],
[1, 0, 2, 0, 2, 1],
[1, 0, 0, 3, 2, 0],
[1, 0, 0, 4, 0, 1]]
TF = self.model._TFIDFModel__convert_tocsr(self.data.item_tags)
self.assertTrue(np.array_equal(TF.todense(), TF_expected))
def test_extract(self):
# test fact extraction without a threshold
DF_expected = [[1, 1, 1, 1, 1, 1],
[1, 0, 1, 0, 1, 0],
[1, 0, 1, 0, 1, 1],
[1, 0, 0, 1, 1, 0],
[1, 0, 0, 1, 0, 1]]
DF = self.model._TFIDFModel__extract_facts(self.data.item_tags)
self.assertTrue(np.array_equal(DF.todense(), DF_expected))
def test_extract_threshold(self):
# test preference extraction with a threshold
P_expected = [[1, 0, 0, 1, 0],
[1, 1, 1, 1, 1],
[1, 0, 0, 0, 1]]
P = self.model._TFIDFModel__extract_facts(self.data.ratings, 3.5)
self.assertTrue(np.array_equal(P.todense(), P_expected))
def test_tfidf_profiles(self):
I_expected = np.matrix([[ 0., 0.86991409, 0.27610534, 0.27610534, 0.12061088, 0.27610534],
[ 0., 0. , 0.9940897 , 0. , 0.10856185, 0. ],
[ 0., 0. , 0.83309624, 0. , 0.36392077, 0.41654812],
[ 0., 0. , 0. , 0.96011533, 0.27960428, 0. ],
[ 0., 0. , 0. , 0.9701425 , 0. , 0.24253563]])
self.model.build(self.data)
# test TFIDF item profile extraction
self.assertTrue(stringify_matrix(self.model.I().todense()) ==
stringify_matrix(I_expected))
