def test_trivial_weighted_mean_entity(self):
points = [Point2D(2, 2), Point2D(2, 5), Point2D(6, 5), Point2D(7, 3), Point2D(4, 7)]
entity = LinearRegressionEntity(points)
weighted_entity = LinearRegressionEntity.weighted_mean_entity([entity])
self.assertAlmostEquals(first=entity.slope, second=weighted_entity.slope, delta=1e-6)
self.assertAlmostEquals(first=entity.offset, second=weighted_entity.offset, delta=1e-6)
def test_hard_mahalanobis_distance_sqr(self):
points1 = [Point2D(2, 2), Point2D(2, 5), Point2D(6, 5), Point2D(7, 3), Point2D(4, 7)]
points2 = [Point2D(6, 5), Point2D(7, 4), Point2D(8, 7), Point2D(5, 6), Point2D(5, 4)]
entity1 = LinearRegressionEntity(points1)
entity2 = LinearRegressionEntity(points2)
d = LinearRegressionEntity.mahalanobis_distance_sqr(entity1, entity2)
self.assertAlmostEqual(first=d, second=0.543724758, delta=1e-6)
def test_can_merge_true(self):
points1 = [Point2D(6, 5), Point2D(7, 4), Point2D(8, 7), Point2D(5, 6), Point2D(5, 4)]
entity1 = LinearRegressionEntity(points1)
coordinator1 = LinearRegressionCoordinator(entity1, 2)
points2 = [Point2D(6, 5), Point2D(7, 4), Point2D(8, 7), Point2D(5, 6), Point2D(5, 4)]
entity2 = LinearRegressionEntity(points2)
coordinator2 = LinearRegressionCoordinator(entity2, 5)
self.assertTrue(coordinator2.can_merge(coordinator1))
def test_init_weighted_mean_entity(self):
points1 = [Point2D(2, 2), Point2D(2, 5), Point2D(6, 5), Point2D(7, 3), Point2D(4, 7)]
points2 = [Point2D(6, 5), Point2D(7, 4), Point2D(8, 7), Point2D(5, 6), Point2D(5, 4)]
entity1 = LinearRegressionEntity(points1)
entity2 = LinearRegressionEntity(points2)
weighted_mean_entity = LinearRegressionEntity.weighted_mean_entity([entity1, entity2])
self.assertIsNotNone(weighted_mean_entity.slope)
self.assertIsNotNone(weighted_mean_entity.offset)
self.assertIsNotNone(weighted_mean_entity.covariance)
def test_merge_same(self):
points1 = [Point2D(6, 5), Point2D(7, 4), Point2D(8, 7), Point2D(5, 6), Point2D(5, 4)]
entity1 = LinearRegressionEntity(points1)
coordinator1 = LinearRegressionCoordinator(entity1, 2)
points2 = [Point2D(6, 5), Point2D(7, 4), Point2D(8, 7), Point2D(5, 6), Point2D(5, 4)]
entity2 = LinearRegressionEntity(points2)
coordinator2 = LinearRegressionCoordinator(entity2, 2)
coordinator = coordinator2.merge(coordinator1)
self.assertEqual(first=coordinator.start_index, second=2)
self.assertEqual(first=len(coordinator.entity.points), second=5)
def test_init(self):
points = [Point2D(6, 5), Point2D(7, 4), Point2D(8, 7), Point2D(5, 6), Point2D(5, 4)]
entity = LinearRegressionEntity(points)
coordinator = LinearRegressionCoordinator(entity, 10)
self.assertEqual(first=entity, second=coordinator.entity)
self.assertEqual(first=coordinator.start_index, second=10)
def test_linear_entity_init(self):
points = [Point2D(0, 0), Point2D(1, 1), Point2D(2, 2)]
entity = LinearRegressionEntity(points)
self.assertEqual(entity.points, points)
self.assertAlmostEqual(first=entity._slope, second=math.tan(math.pi/4.0), delta=1e-6)
self.assertAlmostEqual(first=entity._offset, second=0, delta=1e-6)
self.assertIsNotNone(entity.covariance)
def test_singular_mahalanobis_distance_sqr(self):
points = [Point2D(0, 0), Point2D(1, 1), Point2D(2, 2)]
entity = LinearRegressionEntity(points)
d = LinearRegressionEntity.mahalanobis_distance_sqr(entity, entity)
self.assertAlmostEqual(first=d, second=0.0, delta=1e-6)