def test_art_handle_errors(self):
with self.assertRaises(ValueError):
# Invalid input data dimension
artnet = algorithms.ART1(step=0.4,
rho=0.1,
n_clusters=3,
verbose=False)
artnet.predict(np.array([0.5]))
with self.assertRaises(ValueError):
# Non-binary input values
artnet = algorithms.ART1(step=0.4,
rho=0.1,
n_clusters=3,
verbose=False)
artnet.predict(np.array([[0.5]]))
with self.assertRaises(ValueError):
# Invalid data size for second input
artnet = algorithms.ART1(step=0.4,
rho=0.1,
n_clusters=3,
verbose=False)
artnet.predict(np.array([[0.5]]))
artnet.predict(np.array([[0.5, 1]]))
def test_simple_art1(self):
ann = algorithms.ART1(step=2, rho=0.7, n_clusters=2, verbose=False)
classes = ann.predict(data)
for answer, result in zip([0, 1, 1], classes):
self.assertEqual(result, answer)
self.assertPickledNetwork(ann, data)
def test_art1_on_real_problem(self):
data = pd.DataFrame(lenses)
encoder = preprocessing.OneHotEncoder()
enc_data = encoder.fit_transform(data.values[:, 1:]).toarray()
artnet = algorithms.ART1(step=1.5,
rho=0.7,
n_clusters=3,
verbose=False)
classes = artnet.predict(enc_data)
unique_classes = list(np.sort(np.unique(classes)))
self.assertEqual(unique_classes, [0, 1, 2])
def test_art1_on_real_problem(self):
path_to_data = os.path.join(BASEDIR, '..', 'data', 'lenses.csv')
data = pd.read_csv(path_to_data, index_col=[0], sep=' ', header=None)
encoder = preprocessing.OneHotEncoder()
enc_data = encoder.fit_transform(data.values[:, 1:]).toarray()
artnet = algorithms.ART1(step=1.5,
rho=0.7,
n_clusters=3,
verbose=False)
classes = artnet.predict(enc_data)
self.assertEqual(list(np.sort(np.unique(classes))), [0, 1, 2])
import numpy as np from neupy import algorithms, environment environment.reproducible() file_name = 'dosen' data = np.genfromtxt(file_name + '.csv', delimiter=',') artnet = algorithms.ART1(step=2, rho=0.7, n_clusters=2, verbose=False) artnet.predict(data)