def test_eps_neighborhood(self):
distances = CondensedMatrix([
1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 2., 1., 1., 1., 1.,
1., 3., 1., 1., 1., 1., 2., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.,
2., 1.
])
dbscan_alg = DBSCANAlgorithm(distances)
self.assertItemsEqual(
dbscan_alg._DBSCANAlgorithm__eps_neighborhood(6, 3),
[0, 1, 2, 3, 4, 5, 7, 8])
self.assertItemsEqual(
dbscan_alg._DBSCANAlgorithm__eps_neighborhood(6, 2),
[0, 1, 3, 4, 5, 7, 8])
self.assertItemsEqual(
dbscan_alg._DBSCANAlgorithm__eps_neighborhood(6, 1), [0, 4, 5, 7])
self.assertItemsEqual(
dbscan_alg._DBSCANAlgorithm__eps_neighborhood(6, 1.), [0, 4, 5, 7])
self.assertItemsEqual(distances.element_neighbors_within_radius(6, 3),
[0, 1, 2, 3, 4, 5, 7, 8])
self.assertItemsEqual(distances.element_neighbors_within_radius(6, 2),
[0, 1, 3, 4, 5, 7, 8])
self.assertItemsEqual(distances.element_neighbors_within_radius(6, 1),
[0, 4, 5, 7])
self.assertItemsEqual(distances.element_neighbors_within_radius(6, 1.),
[0, 4, 5, 7])
def test_dbscan_regression_mini(self):
distances = CondensedMatrix([
12.36931688, 5.83095189, 9.43398113, 12.52996409, 15.65247584,
17.4642492, 9.21954446, 4.47213595, 3.16227766, 4.47213595,
5.65685425, 5., 8.06225775, 11.18033989, 13.15294644, 3.16227766,
6.32455532, 8.24621125, 3.16227766, 5.09901951, 2.
])
dbscan_alg = DBSCANAlgorithm(distances)
clustering = dbscan_alg.perform_clustering(kwargs={
"eps": 4.0,
"minpts": 3
})
self.assertItemsEqual(clustering.gen_class_list(starts_with=1),
[-1, 1, -1, 1, 1, 1]) #[0, 1, 0, 1, 1, 1, 0]
def test_dbscan(self):
distances = CondensedMatrix([0., 0., 2., 2., 0., 2., 2., 2., 2., 0.])
dbscan_alg = DBSCANAlgorithm(distances)
eps = 1.0
minpts = 2
clustering = dbscan_alg.perform_clustering(kwargs={
"eps": eps,
"minpts": minpts
})
expected = [1, 1, 1]
self.assertItemsEqual(clustering.gen_class_list(starts_with=1),
expected)
dbscan_alg.element_class = [PointClassType.UNCLASSIFIED] * 5
eps = 1.0
minpts = 1
clustering = dbscan_alg.perform_clustering(kwargs={
"eps": eps,
"minpts": minpts
})
self.assertItemsEqual(clustering.gen_class_list(starts_with=1),
[1, 1, 1, 2, 2])
def test_seed_expansion(self):
"""
Graph:
1--2 3--4
\ /
0
Adj. list:
1 1 0 0 0
1 0 0 0
0 0 0
0 0
1
Inv. distance adjacency list (two nodes are connected if their distance is
lower than x)
0 0 x x x
0 x x x
x x x
x x
0
"""
distances = CondensedMatrix([0., 0., 2., 2., 0., 2., 2., 2., 2., 0.])
N = distances.row_length
dbscan_alg = DBSCANAlgorithm(distances)
eps = 1.0
minpts = 2
elements_class = [PointClassType.UNCLASSIFIED] * N
dbscan_alg._DBSCANAlgorithm__seed_expansion(1, eps, minpts, [2],
elements_class)
expected_classes = [
1, 1, 1, PointClassType.UNCLASSIFIED, PointClassType.UNCLASSIFIED
]
self.assertItemsEqual(expected_classes, elements_class)
elements_class = [PointClassType.UNCLASSIFIED] * N
dbscan_alg._DBSCANAlgorithm__seed_expansion(2, eps, minpts, [3],
elements_class)
expected_classes = [
PointClassType.UNCLASSIFIED, PointClassType.UNCLASSIFIED,
PointClassType.UNCLASSIFIED, PointClassType.UNCLASSIFIED,
PointClassType.UNCLASSIFIED
]
self.assertItemsEqual(expected_classes, elements_class)
minpts = 1
dbscan_alg._DBSCANAlgorithm__seed_expansion(2, eps, minpts, [3],
elements_class)
expected_classes = [
PointClassType.UNCLASSIFIED, PointClassType.UNCLASSIFIED,
PointClassType.UNCLASSIFIED, 2, 2
]
self.assertItemsEqual(expected_classes, elements_class)
elements_class = [PointClassType.UNCLASSIFIED] * N
dbscan_alg._DBSCANAlgorithm__seed_expansion(1, eps, minpts, [2, 4],
elements_class)
expected_classes = [1, 1, 1, 1, 1]
self.assertItemsEqual(expected_classes, elements_class)
