class TestClassifiers(unittest.TestCase):
'''
Integrations tests for classifier objective functions. Checks to see that at least
one iteration can be executed on the given dataset and algorithm.
To increase test speed:
- iterations set to 1
- num_cv_folds=2
'''
iters=1
folds=2
x, y = clf_prep('pima-indians-diabetes.csv')
print('Data prepped')
def test_objective01_xgboost(self):
automator=MLAutomator(self.x, self.y, iterations=self.iters, specific_algos=['01'], num_cv_folds=self.folds)
automator.find_best_algorithm()
self.assertEqual(automator.best_algo, 'xgboost_classifier')
def test_objective02_sgd_classifier(self):
automator=MLAutomator(self.x, self.y, iterations=self.iters, specific_algos=['02'], num_cv_folds=self.folds)
automator.find_best_algorithm()
self.assertEqual(automator.best_algo, 'SGDClassifier')
def test_objective03_sgd_classifier(self):
automator=MLAutomator(self.x, self.y, iterations=self.iters, specific_algos=['03'], num_cv_folds=self.folds)
automator.find_best_algorithm()
self.assertEqual(automator.best_algo, 'RandomForestClassifier')
def test_objective04_bag_of_svc(self):
automator=MLAutomator(self.x, self.y, iterations=self.iters, specific_algos=['04'], num_cv_folds=self.folds)
automator.find_best_algorithm()
self.assertEqual(automator.best_algo, 'SVC')
def test_objective05_naive_bayes(self):
automator=MLAutomator(self.x, self.y, iterations=self.iters, specific_algos=['05'], num_cv_folds=self.folds)
automator.find_best_algorithm()
self.assertEqual(automator.best_algo, 'GaussianNB')
def test_objective06_logistic_regression(self):
automator=MLAutomator(self.x, self.y, iterations=self.iters, specific_algos=['06'], num_cv_folds=self.folds)
automator.find_best_algorithm()
self.assertEqual(automator.best_algo, 'LogisticRegression')
def test_objective07_knn(self):
automator=MLAutomator(self.x, self.y, iterations=self.iters, specific_algos=['07'], num_cv_folds=self.folds)
automator.find_best_algorithm()
self.assertEqual(automator.best_algo, 'KNeighborClassifier')
class TestMLAutomator(unittest.TestCase):
directory = mkdtemp()
x, y = clf_prep('pima-indians-diabetes.csv')
automator = MLAutomator(x, y, iterations=2, specific_algos=['01'])
automator.find_best_algorithm()
def test_fit_best_pipeline(self):
self.automator.fit_best_pipeline()
self.assertIsNotNone(self.automator.best_pipeline)
print(self.automator.best_pipeline)
def test_model_dump(self):
self.automator.save_best_pipeline(self.directory)
def test_model_load(self):
self.automator.load_best_pipeline(filename=self.directory +
'/pipeline.joblib')
class TestMLAutomator(unittest.TestCase):
x, y = clf_prep('pima-indians-diabetes.csv')
automator = MLAutomator(x, y)
def test_automator_initialization(self):
'''
Test that all class properties are being initialized properly.
'''
self.assertEqual(self.automator.best, 0)
self.assertEqual(self.automator.count, 0)
self.assertEqual(self.automator.start_time, None)
self.assertEqual(self.automator.objective, None)
self.assertEqual(self.automator.keys, None)
self.assertEqual(self.automator.master_results, [])
self.assertEqual(self.automator.type, 'classifier')
self.assertEqual(self.automator.score_metric, 'accuracy')
self.assertEqual(self.automator.iterations, 25)
self.assertEqual(self.automator.num_cv_folds, 10)
self.assertEqual(self.automator.repeats, 1)
def test_get_obj_key_list(self):
self.assertIsNotNone(classifiers().keys())
self.assertIsNotNone(regressors().keys())
def test_get_keys(self):
for key in ALGORITHM_KEYS.keys():
self.assertIsNotNone(get_keys(key))
print(get_keys(key))
def test_get_space_regressors(self):
for key in regressors().keys():
self.assertIsNotNone(get_space(self.automator, key))
def test_get_space_classifiers(self):
for key in classifiers().keys():
self.assertIsNotNone(get_space(self.automator, key))
def test_user_feedback_went_best_space_not_evaluated(self):
self.assertIsNone(self.automator.print_best_space())
self.assertIsNone(self.automator.save_best_model())
self.assertIsNone(self.automator.fit_best_model())
from data.utilities import clf_prep
from mlautomator.mlautomator import MLAutomator
if __name__ == '__main__':
x, y = clf_prep('boston_housing.csv')
automator = MLAutomator(x, y, iterations=20)
automator.find_best_algorithm()
automator.print_best_space()
from data.utilities import clf_prep
from mlautomator.mlautomator import MLAutomator
if __name__=='__main__':
x,y=clf_prep('GOLD_D.csv')
automator=MLAutomator(x, y, iterations=200, specific_algos=['01'], score_metric='neg_log_loss')
automator.find_best_algorithm()
automator.print_best_space()
#automator.fit_best_model()
from data.utilities import clf_prep
from mlautomator.mlautomator import MLAutomator
if __name__=='__main__':
x,y=clf_prep('pima-indians-diabetes.csv')
automator=MLAutomator(x,y,iterations=30)
#automator.find_best_algorithm()
automator.print_best_space()
print(automator)