def test_returns_linear_regression_params(self):
too_short = TooShort()
lr_params = {
'normalize': [True, False]
}
result = too_short.get_param_grid(LinearRegression)
self.assertEqual(result, lr_params)
def test_returns_regression_models_many_samples(self):
too_short = TooShort(prediction_type="regression")
y = np.random.choice([0, 1, 2, 3, 4], 110000)
too_short.set_attributes(y_train=y)
result = too_short.choose_models()
self.assertIn(LinearRegression, result)
self.assertIn(SGDRegressor, result)
def test_returns_classification_models_many_samples(self):
too_short = TooShort(prediction_type="classification")
y = np.random.choice([0, 1, 2, 3, 4], 110000)
too_short.set_attributes(y_train=y)
result = too_short.choose_models()
self.assertIn(SVC, result)
self.assertIn(SGDClassifier, result)
def testBasicOversamplingNoDfNoUndersampling(self):
too_short = TooShort()
X, y = make_classification(n_samples=10000, n_features=2, n_redundant=0,
n_clusters_per_class=1, weights=[0.75], flip_y=0, random_state=1)
too_short.set_attributes(X_train=X, y_train=y)
os_X, os_y = too_short.oversample()
count = Counter(os_y)
self.assertTrue(count[0] == 7500)
self.assertTrue(count[1] == 7500)
def testBasicFeatureSelection(self):
X, y = get_iris()
too_short = TooShort(X, y, prediction_type="classification")
X_train, X_test = too_short.preproc(
standard_scale=['alcohol', 'malic_acid', 'ash', 'alcalinity_of_ash', 'magnesium',
'total_phenols', 'flavanoids', 'nonflavanoid_phenols',
'proanthocyanins', 'color_intensity', 'hue',
'od280/od315_of_diluted_wines', 'proline'])
X_train_filtered, X_test_filtered = too_short.select_features()
self.assertTrue(len(X_train.columns) > len(X_train_filtered.columns))
self.assertTrue(len(X_test.columns) > len(X_test_filtered.columns))
def testRegressionSmallEndToEnd(self):
X, y = get_boston()
too_short = TooShort(X, y, prediction_type="regression")
result = too_short.preproc(
standard_scale=['CRIM', 'ZN', 'INDUS', 'CHAS', 'NOX', 'RM', 'AGE', 'DIS', 'RAD', 'TAX',
'PTRATIO', 'B', 'LSTAT'])
models = too_short.choose_models()
result = too_short.search()
model_keys = result.keys()
self.assertIn('Ridge', model_keys)
def testCatSmallEndToEnd(self):
X, y = get_iris()
too_short = TooShort(X, y, prediction_type="classification")
result = too_short.preproc(
standard_scale=['alcohol', 'malic_acid', 'ash', 'alcalinity_of_ash', 'magnesium',
'total_phenols', 'flavanoids', 'nonflavanoid_phenols',
'proanthocyanins', 'color_intensity', 'hue',
'od280/od315_of_diluted_wines', 'proline'])
models = too_short.choose_models()
result = too_short.search()
model_keys = result.keys()
self.assertIn('SVC', model_keys)
def test_standard_scaled(self):
X, y = get_wine()
X['A_FAKE_CAT'] = np.random.randint(4, size=len(y))
X['B_FAKE_CAT'] = np.random.randint(4, size=len(y))
X['C_FAKE_CAT'] = np.random.choice(['SWEET', 'SOUR', 'TART'], len(y))
X['D_FAKE_LABEL_CAT'] = np.random.choice(
['BAD', 'OK', 'GOOD', 'GREAT'], len(y))
too_short = TooShort(X, y)
result = too_short.preproc(
standard_scale=['alcohol', 'malic_acid', 'ash', 'alcalinity_of_ash', 'magnesium',
'total_phenols', 'flavanoids', 'nonflavanoid_phenols',
'proanthocyanins', 'color_intensity', 'hue',
'od280/od315_of_diluted_wines', 'proline'])
result_df = result[0]
self.assertAlmostEqual(
result_df['alcohol'].mean(), result_df['malic_acid'].mean())
def test_does_not_alter_original_df(self):
X, y = get_wine()
X['A_FAKE_CAT'] = np.random.randint(4, size=len(y))
X['B_FAKE_CAT'] = np.random.randint(4, size=len(y))
X['C_FAKE_CAT'] = np.random.choice(['SWEET', 'SOUR', 'TART'], len(y))
X['D_FAKE_LABEL_CAT'] = np.random.choice(
['BAD', 'OK', 'GOOD', 'GREAT'], len(y))
X_copy = X.copy()
too_short = TooShort(X, y)
too_short.preproc(OHE=np.array(
['A_FAKE_CAT', 'B_FAKE_CAT', 'C_FAKE_CAT']),
label_encode={
'D_FAKE_LABEL_CAT': ['BAD', 'OK', 'GOOD', 'GREAT']
},
standard_scale=['alcohol', 'malic_acid', 'ash', 'alcalinity_of_ash', 'magnesium',
'total_phenols', 'flavanoids', 'nonflavanoid_phenols',
'proanthocyanins', 'color_intensity', 'hue',
'od280/od315_of_diluted_wines', 'proline'])
assert_frame_equal(X, X_copy)
def test_create_ohe(self):
X, y = get_wine()
X['A_FAKE_CAT'] = np.random.randint(4, size=len(y))
X['B_FAKE_CAT'] = np.random.randint(4, size=len(y))
X['C_FAKE_CAT'] = np.random.choice(['SWEET', 'SOUR', 'TART'], len(y))
X['D_FAKE_LABEL_CAT'] = np.random.choice(
['BAD', 'OK', 'GOOD', 'GREAT'], len(y))
too_short = TooShort(X, y)
X_train, X_test = too_short.preproc(OHE=np.array(
['A_FAKE_CAT', 'B_FAKE_CAT', 'C_FAKE_CAT']),
label_encode={
'D_FAKE_LABEL_CAT': ['BAD', 'OK', 'GOOD', 'GREAT']
},
standard_scale=['alcohol', 'malic_acid', 'ash', 'alcalinity_of_ash', 'magnesium',
'total_phenols', 'flavanoids', 'nonflavanoid_phenols',
'proanthocyanins', 'color_intensity', 'hue',
'od280/od315_of_diluted_wines', 'proline'])
result_df = X_train
self.assertCountEqual(result_df.columns[0:11], ['A_FAKE_CAT_0', 'A_FAKE_CAT_1', 'A_FAKE_CAT_2', 'A_FAKE_CAT_3',
'B_FAKE_CAT_0', 'B_FAKE_CAT_1', 'B_FAKE_CAT_2', 'B_FAKE_CAT_3',
'C_FAKE_CAT_SOUR', 'C_FAKE_CAT_SWEET', 'C_FAKE_CAT_TART'])
self.assertFalse('A_FAKE_CAT' in result_df.columns)
self.assertIn(result_df['A_FAKE_CAT_0'][0], [0.0, 1.0])
def test_basic_custom_grid_search(self):
X, y = get_iris()
too_short = TooShort(X, y)
result = too_short.preproc(
standard_scale=too_short.X_train.columns)
too_short.set_attributes(models=[KNeighborsClassifier])
result = too_short.search()
model_keys = result.keys()
self.assertEqual(len(model_keys), 1)
self.assertIn('KNeighborsClassifier', model_keys)
def test_returns_classification_models_small_samples(self):
X, y = get_iris()
too_short = TooShort(X, y, prediction_type="classification")
result = too_short.choose_models()
self.assertIn(SVC, result)
self.assertNotIn(SGDClassifier, result)
def test_returns_regression_models_small_samples(self):
X, y = get_iris()
too_short = TooShort(X, y, prediction_type="regression")
result = too_short.choose_models()
self.assertIn(LinearRegression, result)
self.assertNotIn(SGDRegressor, result)