def test_sgd(self):
r_matrix = np.array([[5, 3, 0, 1], [4, 0, 0, 1]])
mf = matrixfactorization.MatrixFactorization(r_matrix)
mf.sgd()
assert (mf.biase_users.tolist()
== [0.007484458726407961, -0.035414091521230456]
and mf.item_latfact_matrix.tolist() == [[
-0.032641133006462326, -0.05555365174071124
], [0.7854491962952435, 0.38496207636014385
], [-0.23473719296747605, 0.27128002179298233
], [-0.33370623672354505, -0.33482884637651444]])
def test_full_matrix(self):
r_matrix = np.array([[5, 3, 0, 1], [4, 0, 0, 1]])
mf = matrixfactorization.MatrixFactorization(r_matrix)
assert mf.full_matrix().tolist() == [[
2.7790163718826255, 2.9695771323896034, 2.7229471856900074,
2.758551918034064
],
[
2.672935991777994,
3.347916011292537,
2.9305654916433834,
2.5476323380768116
]]
def test_train(self):
r_matrix = np.array([
[5, 3, 0, 1],
[4, 0, 0, 1],
[1, 1, 0, 5],
[1, 0, 0, 4],
[0, 1, 5, 4],
])
mf = matrixfactorization.MatrixFactorization(r_matrix)
assert mf.train() == [
(0, 5.468596914045281), (1, 4.954645307979005),
(2, 4.300804333132003), (3, 3.3569039342211373),
(4, 2.126600667778168), (5, 1.2509019281100942),
(6, 0.8696031656138972), (7, 0.6818809684270286),
(8, 0.5669838258059611), (9, 0.48277705082712996),
(10, 0.4003761329325773), (11, 0.3465686840777396),
(12, 0.2835041378952923), (13, 0.24552957914749302),
(14, 0.21526948953431074), (15, 0.18468907691426226),
(16, 0.1628286226059519), (17, 0.13394758593152065),
(18, 0.11881052024819237), (19, 0.10858704529104617)
]
def test_get_rating(self):
r_matrix = np.array([[5, 3, 0, 1], [4, 0, 0, 1]])
mf = matrixfactorization.MatrixFactorization(r_matrix)
assert mf.get_rating(1, 0) == 2.672935991777994
def test_mes(self):
r_matrix = np.array([[5, 3, 0, 1], [4, 0, 0, 1]])
mf = matrixfactorization.MatrixFactorization(r_matrix)
assert mf.mse() == 3.4903385812037913