def test_nominal_info_gain(self):
x, y, t, feature_names, name, dataset_type = load.breast_cancer()
actual = np.array([measures.info_gain(x[:, i], y, t[i], True)[0] for i in range(len(x[0]))])
expected = np.array(
[0.363243, 0.588919, 0.569025, 0.375882, 0.494187, 0.520238, 0.490311, 0.457238, 0.199398])
np.testing.assert_allclose(expected, actual, rtol=1e-05)
def dt_multiple_datasets(test_cases, args):
print "DECISION TREE"
print args
for num_test in test_cases:
if num_test == 1:
decision_tree_ca(load.breast_cancer(), args)
decision_tree_ca(load.car(), args)
elif num_test == 2:
decision_tree_ca(load.segmentation(), args)
decision_tree_ca(load.iris(), args)
decision_tree_ca(load.wine(), args)
elif num_test == 3:
decision_tree_ca(load.bank(), args)
decision_tree_ca(load.lymphography(), args)
def clasifier_comparison(test_cases, args):
print "Comparison of DECISION TREE, RANDOM FOREST"
print args
for num_test in test_cases:
if num_test == 1:
compare(load.breast_cancer(), args)
compare(load.car(), args)
elif num_test == 2:
compare(load.segmentation(), args)
compare(load.iris(), args)
compare(load.wine(), args)
if num_test == 3:
compare(load.bank(), args)
compare(load.lymphography(), args)
