def test_predict_different_inputs(self): lmsnet = algorithms.LMS((1, 1), verbose=False) data = np.array([[1, 2, 3]]).T target = np.array([[1, 1, 1]]).T lmsnet.train(data, target) self.assertInvalidVectorPred(lmsnet, data.ravel(), target, decimal=2)
def test_lms_output_data_type(self): input_data = np.array([[1, 0], [2, 2], [3, 3], [0, 0]]) target_data = np.array([[1], [0], [0], [1]]) lmsnet = algorithms.LMS((2, 1), step=0.1) lmsnet.train(input_data, target_data, epochs=200) predicted = lmsnet.predict(np.array([[4, 4], [0, 0]])) self.assertTrue(np.issubdtype(predicted.dtype, np.integer))
def test_train(self): input_data = np.array([[1, 0], [2, 2], [3, 3], [0, 0]]) target_data = np.array([[1], [0], [0], [1]]) network = algorithms.LMS((2, 1), step=0.2, verbose=False) network.train(input_data, target_data, epochs=100) predicted_result = network.predict(np.array([[4, 4], [0, 0]])) np.testing.assert_array_almost_equal(predicted_result, np.array([[0, 1]]).T)
def test_train_different_inputs(self): self.assertInvalidVectorTrain(algorithms.LMS((1, 1), verbose=False), np.array([1, 2, 3]), np.array([1, 2, 3]))