def test_get_delta(self):
feature_selection = fs.FeatureSelection(None, None, None, None, None,
['IG', '3', '5'])
self.assertEqual(5, feature_selection.get_delta())
feature_selection = fs.FeatureSelection(None, None, None, None, None,
['IG', '3'])
self.assertEqual(0, feature_selection.get_delta())
def test_get_fold(self):
feature_selection = fs.FeatureSelection(['trainingn'] * 4, None, None,
None, None, ['IG', '3', '5'])
self.assertEqual(3, feature_selection.get_fold())
feature_selection = fs.FeatureSelection(['trainingn'] * 14, None, None,
None, None, ['IG'])
self.assertEqual(10, feature_selection.get_fold())
def test_find_attributes_by_ranking(self):
path = datasetsDir(self) + 'minigolf' + SEP + 'weather'
_training = training(path)
attributes, klass = metadata(path)
_test = test(path)
_gold = gold(path)
feature_selection = fs.FeatureSelection(_training, attributes, klass,
_test, _gold, ['IG', '3'])
ig_for_attr1 = information_gain(attributes[0], klass, _training)
self.assertAlmostEqual(0.324409, ig_for_attr1, 6)
self.assertEqual('outlook', attributes[0].name)
ig_for_attr2 = information_gain(attributes[1], klass, _training)
self.assertAlmostEqual(0.102187, ig_for_attr2, 6)
self.assertEqual('temperature', attributes[1].name)
ig_for_attr3 = information_gain(attributes[2], klass, _training)
self.assertAlmostEqual(0.091091, ig_for_attr3, 6)
self.assertEqual('humidity', attributes[2].name)
ig_for_attr4 = information_gain(attributes[3], klass, _training)
self.assertAlmostEqual(0.072780, ig_for_attr4, 6)
self.assertEqual('windy', attributes[3].name)
attributes_to_remove = feature_selection.find_attributes_by_ranking(
'information_gain', 3)
self.assertEqual(1, len(attributes_to_remove))
self.assertEqual('windy', attributes_to_remove[0].name)
def test_invert_attrbute_selection(self):
path = datasetsDir(self) + 'numerical' + SEP + 'person'
_attributes, _klass = metadata(path)
feature_selection = fs.FeatureSelection(None, _attributes, None, None, None, ['IG'])
unselected = feature_selection.invert_attribute_selection([_attributes[0], _attributes[1]])
self.assertEqual(len(_attributes) - 2, len(unselected))
self.assertEqual([_attributes[2], _attributes[3], _attributes[4], _attributes[5], _attributes[6], _attributes[7]], unselected)
def test_backward_select(self):
path = datasetsDir(self) + 'minigolf' + SEP + 'weather'
_training = training(path)
_attributes, _klass = metadata(path)
_test = test(path)
_gold = gold(path)
verify_training = copy.deepcopy(_training)
verify_attributes = copy.deepcopy(_attributes)
feat_sel = fs.FeatureSelection(_training, _attributes, _klass, _test, _gold, ['1R', '4', '0.1'])
feat_sel.backward_elimination()
self.assertEqual(3, len(_attributes))
self.verify_number_of_attributes(_training, 3)
self.verify_number_of_attributes(_test, 3)
self.verify_number_of_attributes(_gold, 3)
#verification
#level 0
avg_acc = feat_sel.avg_accuracy_by_cross_validation(verify_training.cross_validation_datasets(4), 4, verify_attributes)
self.assertAlmostEqual(0.5416666, avg_acc, 6)
verification_cv_datasets = verify_training.cross_validation_datasets(4)
accuracies = {}
for attribute in verify_attributes:
attributes = verify_attributes[:]
attributes.remove(attribute)
accuracies[(attributes[0].name,attributes[1].name,attributes[2].name)] = feat_sel.avg_accuracy_by_cross_validation(verification_cv_datasets, 4, attr.Attributes(attributes))
# {('outlook', 'humidity', 'windy'): 0.54166666666666663,
# ('outlook', 'temperature', 'windy'): 0.54166666666666663,
# ('temperature', 'humidity', 'windy'): 0.29166666666666663,
# ('outlook', 'temperature', 'humidity'): 0.79166666666666663}
self.assertAlmostEqual(0.5416666, accuracies[('outlook', 'humidity', 'windy')], 6)
self.assertAlmostEqual(0.5416666, accuracies[('outlook', 'temperature', 'windy')], 6)
self.assertAlmostEqual(0.2916666, accuracies[('temperature', 'humidity', 'windy')], 6)
self.assertAlmostEqual(0.7916666, accuracies[('outlook', 'temperature', 'humidity')], 6)
#
#('outlook', 'temperature', 'humidity') selected
accuracies = {}
for each in verify_attributes:
if each.name == 'windy':
windy = each
verify_attributes.remove(windy)
for attribute in verify_attributes:
attributes = verify_attributes[:]
attributes.remove(attribute)
accuracies[(attributes[0].name,attributes[1].name)] = feat_sel.avg_accuracy_by_cross_validation(verification_cv_datasets, 4, attr.Attributes(attributes))
self.assertAlmostEqual(0.7916666, accuracies[('outlook','humidity')], 6)
self.assertAlmostEqual(0.7916666, accuracies['outlook', 'temperature'], 6)
self.assertAlmostEqual(0.4166666, accuracies[('temperature','humidity')], 6)
def test_cannot_perform_rank_select_on_cont_attrs(self):
path = datasetsDir(self) + 'numerical' + SEP + 'person'
_training = training(path)
attributes, klass = metadata(path)
_test = test(path)
feature_selection = fs.FeatureSelection(_training, attributes, klass, _test, None, ['IG','2'])
try:
feature_selection.by_rank()
self.fail('should throw error as path points to continuous attributes')
except inv.InvalidDataError:
pass
def test_forward_select(self):
path = datasetsDir(self) + 'minigolf' + SEP + 'weather'
_training = training(path)
_attributes, _klass = metadata(path)
_test = test(path)
_gold = gold(path)
verify_training = copy.deepcopy(_training)
verify_attributes = copy.deepcopy(_attributes)
feat_sel = fs.FeatureSelection(_training, _attributes, _klass, _test,
_gold, ['1R', '4', '0.1'])
feat_sel.forward_selection()
self.assertEqual(1, len(_attributes))
self.assertEqual('outlook', _attributes[0].name)
self.verify_number_of_attributes(_training, 1)
self.verify_number_of_attributes(_test, 1)
self.verify_number_of_attributes(_gold, 1)
#verification
verification_cv_datasets = verify_training.cross_validation_datasets(4)
accuracies = {}
for attribute in verify_attributes:
accuracies[
attribute.name] = feat_sel.avg_accuracy_by_cross_validation(
verification_cv_datasets, 4, attr.Attributes([attribute]))
#'windy': 0.41666666666666663, 'outlook': 0.79166666666666663, 'temperature': 0.41666666666666663, 'humidity': 0.54166666666666663
self.assertAlmostEqual(0.4166666, accuracies['windy'], 6)
self.assertAlmostEqual(0.79166666, accuracies['outlook'], 6)
self.assertAlmostEqual(0.4166666, accuracies['temperature'], 6)
self.assertAlmostEqual(0.5416666, accuracies['humidity'], 6)
#outlook selected
accuracies = {}
for each in verify_attributes:
if each.name == 'outlook':
outlook = each
verify_attributes.remove(outlook)
for attribute in verify_attributes:
accuracies[(
'outlook',
attribute.name)] = feat_sel.avg_accuracy_by_cross_validation(
verification_cv_datasets, 4,
attr.Attributes([outlook, attribute]))
#{('outlook', 'humidity'): 0.79166666666666663, ('outlook', 'temperature'): 0.79166666666666663, ('outlook', 'windy'): 0.54166666666666663}
self.assertAlmostEqual(0.7916666, accuracies[('outlook', 'humidity')],
6)
self.assertAlmostEqual(0.7916666, accuracies['outlook', 'temperature'],
6)
self.assertAlmostEqual(0.5416666, accuracies[('outlook', 'windy')], 6)