def setUp(self):
self.n_queries = 2
self.n_neighbors = 3
self.n_classifiers = 5
from sklearn.datasets import load_iris
from sklearn.linear_model import LogisticRegression
lr = LogisticRegression()
self.bag = Bagging(lr, pool_size=self.n_classifiers)
self.X, self.y = load_iris(return_X_y=True)
np.random.seed(5)
self.predictions = np.random.randint(low=np.min(self.y),
high=np.max(self.y),
size=(self.n_queries,
self.n_neighbors,
self.n_classifiers))
self.predictions_query = np.random.randint(low=np.min(self.y),
high=np.max(self.y),
size=(self.n_queries,
self.n_classifiers))
self.bag.fit(self.X, self.y)
self.selector = LCA(self.X, self.y, n_neighbors=self.n_neighbors)
class TestBaseSelection(unittest.TestCase):
def setUp(self):
self.true_labels = np.array([1, 1, 0, 0, 1, 0])
self.predictions1 = np.array([1, 0, 0, 0, 0, 0])
self.predictions2 = np.array([0, 1, 0, 0, 1, 1])
self.predictions3 = np.array([1, 0, 0, 0, 0, 1])
from sklearn.datasets import load_iris
from sklearn.linear_model import LogisticRegression
lr = LogisticRegression()
self.bag = Bagging(lr, pool_size=5)
self.X, self.y = load_iris(return_X_y=True)
self.bag.fit(self.X, self.y)
self.selector = BaseSelection(self.X, self.y, n_neighbors=3)
def test_predict_nearest(self):
predictions, actual_values = self.selector.predict_nearest(
self.X[0:2, :], self.bag)
self.assertEqual(predictions.shape, (2, 3, 1))
self.assertEqual(actual_values.shape, (2, 3))
def setUp(self):
from sklearn.datasets import load_iris
from sklearn.linear_model import LogisticRegression
lr = LogisticRegression()
self.single = Bagging(lr, pool_size=1)
self.classifier = Bagging(lr, pool_size=5)
self.X, self.y = load_iris(return_X_y=True)
def setUp(self):
self.true_labels = np.array([1, 1, 0, 0, 1, 0])
self.predictions1 = np.array([1, 0, 0, 0, 0, 0])
self.predictions2 = np.array([0, 1, 0, 0, 1, 1])
self.predictions3 = np.array([1, 0, 0, 0, 0, 1])
from sklearn.datasets import load_iris
from sklearn.linear_model import LogisticRegression
lr = LogisticRegression()
self.bag = Bagging(lr, pool_size=5)
self.X, self.y = load_iris(return_X_y=True)
self.bag.fit(self.X, self.y)
self.selector = BaseSelection(self.X, self.y, n_neighbors=3)
class test_ola(unittest.TestCase):
def setUp(self):
self.n_queries = 2
self.n_neighbors = 3
self.n_classifiers = 5
from sklearn.datasets import load_iris
from sklearn.linear_model import LogisticRegression
lr = LogisticRegression()
self.bag = Bagging(lr, pool_size=self.n_classifiers)
self.X, self.y = load_iris(return_X_y=True)
np.random.seed(5)
self.predictions = np.random.randint(low=np.min(self.y),
high=np.max(self.y),
size=(self.n_queries,
self.n_neighbors,
self.n_classifiers))
self.predictions_query = np.random.randint(low=np.min(self.y),
high=np.max(self.y),
size=(self.n_queries,
self.n_classifiers))
self.bag.fit(self.X, self.y)
self.selector = LCA(self.X, self.y, n_neighbors=self.n_neighbors)
def test_local_class_accuracy(self):
_, actual_labels = self.selector.predict_nearest(
self.X[:self.n_queries, :], self.bag)
a = self.selector.best_local_class_accuracy(self.predictions,
self.predictions_query,
actual_labels)
self.assertEqual(a.shape[0], self.n_queries)
def test_select_and_predict(self):
predictions = self.selector.select_and_predict(self.X[:3], self.bag)
self.assertEqual(predictions.shape, (3, ))
class TestBagging(TestCase):
def setUp(self):
from sklearn.datasets import load_iris
from sklearn.linear_model import LogisticRegression
lr = LogisticRegression()
self.single = Bagging(lr, pool_size=1)
self.classifier = Bagging(lr, pool_size=5)
self.X, self.y = load_iris(return_X_y=True)
def test_fit(self):
self.single.fit(self.X, self.y)
self.classifier.fit(self.X, self.y)
self.assertTrue(self.classifier.has_been_fit, msg='Classifier fit')
def test_predict(self):
self.classifier.fit(self.X, self.y)
y_pred = self.classifier.predict(self.X)
self.assertEqual(len(y_pred), len(self.X))