def test_feature_scaling():
matrix, dep_vars = cleanup_data(import_data())
matrix_train, matrix_test, dependent_train, dependent_test = training_set(
encode_feature(matrix), encode_feature(dep_vars))
scaled_matrix_train, scaled_matrix_test = feature_scaling(
matrix_train, matrix_test)
assert scaled_matrix_train.shape == matrix_train.shape
def test_training_set():
matrix_train, matrix_test, dependent_train, dependent_test = training_set(
*features_and_dependent_vars(import_data()))
assert matrix_train.shape == (8, 3)
assert matrix_test.shape == (2, 3)
assert dependent_train.shape == (8, 1)
assert dependent_test.shape == (2, 1)
def test_training_and_test_sets():
train_x, test_x, train_y, test_y = dp.training_set(
*dp.features_and_dependent_vars(dp.import_data(DATA_FILE)),
test_size=1 / 3)
assert train_x[0][0] == 2.9
assert test_x[0][0] == 1.5
assert train_y[0][0] == 56642
assert test_y[0][0] == 37731
def test_predict():
train_x, test_x, train_y, test_y = dp.training_set(
*dp.features_and_dependent_vars(dp.import_data(DATA_FILE)),
test_size=1 / 3)
machine = train_the_machine(train_x, train_y)
predicted = predict(machine, train_x)
data, max_error = error(train_y, predicted)
assert max_error < 0.2
assert data['err'].mean() < 0.1
def test_train_the_machine():
train_x, _, train_y, _ = dp.training_set(*dp.features_and_dependent_vars(
dp.import_data(DATA_FILE)),
test_size=1 / 3)
machine = train_the_machine(train_x, train_y)
assert isinstance(machine, LinearRegression)