1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
|
# Test methods with long descriptive names can omit docstrings
# pylint: disable=missing-docstring
import pickle
import unittest
import numpy as np
from Orange.data import Table
from Orange.preprocess import Continuize, Normalize
from Orange.projection import PCA, SparsePCA, IncrementalPCA, TruncatedSVD
from Orange.tests import test_filename
class TestPCA(unittest.TestCase):
@classmethod
def setUpClass(cls):
cls.ionosphere = Table(test_filename('datasets/ionosphere.tab'))
cls.iris = Table('iris')
cls.zoo = Table('zoo')
def test_pca(self):
data = self.ionosphere
self.__pca_test_helper(data, n_com=3, min_xpl_var=0.5)
self.__pca_test_helper(data, n_com=10, min_xpl_var=0.7)
self.__pca_test_helper(data, n_com=32, min_xpl_var=1)
def __pca_test_helper(self, data, n_com, min_xpl_var):
pca = PCA(n_components=n_com)
pca_model = pca(data)
pca_xpl_var = np.sum(pca_model.explained_variance_ratio_)
self.assertGreaterEqual(pca_xpl_var + 1e-6, min_xpl_var)
self.assertEqual(n_com, pca_model.n_components)
self.assertEqual((n_com, data.X.shape[1]), pca_model.components_.shape)
proj = np.dot(data.X - pca_model.mean_, pca_model.components_.T)
np.testing.assert_almost_equal(pca_model(data).X, proj)
def test_sparse_pca(self):
data = self.ionosphere[:100]
self.__sparse_pca_test_helper(data, n_com=3, max_err=1500)
self.__sparse_pca_test_helper(data, n_com=10, max_err=1000)
self.__sparse_pca_test_helper(data, n_com=32, max_err=500)
def __sparse_pca_test_helper(self, data, n_com, max_err):
sparse_pca = SparsePCA(n_components=n_com, ridge_alpha=0.001, random_state=0)
pca_model = sparse_pca(data)
self.assertEqual(n_com, pca_model.n_components)
self.assertEqual((n_com, data.X.shape[1]), pca_model.components_.shape)
self.assertLessEqual(pca_model.error_[-1], max_err)
def test_randomized_pca(self):
data = self.ionosphere
self.__rnd_pca_test_helper(data, n_com=3, min_xpl_var=0.5)
self.__rnd_pca_test_helper(data, n_com=10, min_xpl_var=0.7)
self.__rnd_pca_test_helper(data, n_com=32, min_xpl_var=0.98)
def __rnd_pca_test_helper(self, data, n_com, min_xpl_var):
rnd_pca = PCA(n_components=n_com, svd_solver='randomized')
pca_model = rnd_pca(data)
pca_xpl_var = np.sum(pca_model.explained_variance_ratio_)
self.assertGreaterEqual(pca_xpl_var, min_xpl_var)
self.assertEqual(n_com, pca_model.n_components)
self.assertEqual((n_com, data.X.shape[1]), pca_model.components_.shape)
proj = np.dot(data.X - pca_model.mean_, pca_model.components_.T)
np.testing.assert_almost_equal(pca_model(data).X, proj)
def test_incremental_pca(self):
data = self.ionosphere
self.__ipca_test_helper(data, n_com=3, min_xpl_var=0.49)
self.__ipca_test_helper(data, n_com=32, min_xpl_var=1)
def __ipca_test_helper(self, data, n_com, min_xpl_var):
pca = IncrementalPCA(n_components=n_com)
pca_model = pca(data[::2])
pca_xpl_var = np.sum(pca_model.explained_variance_ratio_)
self.assertGreaterEqual(pca_xpl_var + 1e-6, min_xpl_var)
self.assertEqual(n_com, pca_model.n_components)
self.assertEqual((n_com, data.X.shape[1]), pca_model.components_.shape)
proj = np.dot(data.X - pca_model.mean_, pca_model.components_.T)
np.testing.assert_almost_equal(pca_model(data).X, proj)
pc1_ipca = pca_model.components_[0]
self.assertAlmostEqual(np.linalg.norm(pc1_ipca), 1)
pc1_pca = PCA(n_components=n_com)(data).components_[0]
self.assertAlmostEqual(np.linalg.norm(pc1_pca), 1)
self.assertNotAlmostEqual(abs(pc1_ipca.dot(pc1_pca)), 1, 2)
pc1_ipca = pca_model.partial_fit(data[1::2]).components_[0]
self.assertAlmostEqual(abs(pc1_ipca.dot(pc1_pca)), 1, 4)
def test_truncated_svd(self):
data = self.ionosphere
self.__truncated_svd_test_helper(data, n_components=3, min_variance=0.5)
self.__truncated_svd_test_helper(data, n_components=10, min_variance=0.7)
self.__truncated_svd_test_helper(data, n_components=31, min_variance=0.99)
def __truncated_svd_test_helper(self, data, n_components, min_variance):
model = TruncatedSVD(n_components=n_components)(data)
svd_variance = np.sum(model.explained_variance_ratio_)
self.assertGreaterEqual(svd_variance + 1e-6, min_variance)
self.assertEqual(n_components, model.n_components)
self.assertEqual((n_components, data.X.shape[1]), model.components_.shape)
proj = np.dot(data.X, model.components_.T)
np.testing.assert_almost_equal(model(data).X, proj)
def test_compute_value(self):
iris = self.iris
pca = PCA(n_components=2)(iris)
pca_iris = pca(iris)
pca_iris2 = iris.transform(pca_iris.domain)
np.testing.assert_almost_equal(pca_iris.X, pca_iris2.X)
np.testing.assert_equal(pca_iris.Y, pca_iris2.Y)
pca_iris3 = pickle.loads(pickle.dumps(pca_iris))
np.testing.assert_almost_equal(pca_iris.X, pca_iris3.X)
np.testing.assert_equal(pca_iris.Y, pca_iris3.Y)
def test_transformed_domain_does_not_pickle_data(self):
iris = self.iris
pca = PCA(n_components=2)(iris)
pca_iris = pca(iris)
pca_iris2 = iris.transform(pca_iris.domain)
pca_iris2 = pickle.loads(pickle.dumps(pca_iris))
self.assertIsNone(pca_iris2.domain[0].compute_value.transformed)
def test_chain(self):
zoo_c = Continuize()(self.zoo)
pca = PCA(n_components=3)(zoo_c)(self.zoo)
pca2 = PCA(n_components=3)(zoo_c)(zoo_c)
pp = [Continuize()]
pca3 = PCA(n_components=3, preprocessors=pp)(self.zoo)(self.zoo)
np.testing.assert_almost_equal(pca.X, pca2.X)
np.testing.assert_almost_equal(pca.X, pca3.X)
def test_PCA_scorer(self):
data = self.iris
pca = PCA(preprocessors=[Normalize()])
pca.component = 1
scores = pca.score_data(data)
self.assertEqual(scores.shape[1], len(data.domain.attributes))
self.assertEqual(['petal length', 'petal width'],
sorted([data.domain.attributes[i].name
for i in np.argsort(scores[0])[-2:]]))
self.assertEqual([round(s, 4) for s in scores[0]],
[0.5224, 0.2634, 0.5813, 0.5656])
def test_PCA_scorer_component(self):
pca = PCA()
for i in range(1, len(self.zoo.domain.attributes) + 1):
pca.component = i
scores = pca.score_data(self.zoo)
self.assertEqual(scores.shape,
(pca.component, len(self.zoo.domain.attributes)))
def test_PCA_scorer_all_components(self):
n_attr = len(self.iris.domain.attributes)
pca = PCA()
scores = pca.score_data(self.iris)
self.assertEqual(scores.shape, (n_attr, n_attr))
def test_max_components(self):
d = np.random.RandomState(0).rand(20, 20)
data = Table.from_numpy(None, d)
pca = PCA()(data)
self.assertEqual(len(pca.explained_variance_ratio_), 20)
pca = PCA(n_components=10)(data)
self.assertEqual(len(pca.explained_variance_ratio_), 10)
def test_eq_hash(self):
d = np.random.RandomState(0).rand(20, 20)
data = Table.from_numpy(None, d)
p1 = PCA()(data)
p2 = PCA()(data)
np.testing.assert_equal(p1(data).X, p2(data).X)
# even though results are the same, these transformations
# are different because the PCA object is
self.assertNotEqual(p1, p2)
self.assertNotEqual(p1.domain, p2.domain)
self.assertNotEqual(hash(p1), hash(p2))
self.assertNotEqual(hash(p1.domain), hash(p2.domain))
def test_eq_hash_fake_same_projection(self):
d = np.random.RandomState(0).rand(20, 20)
data = Table.from_numpy(None, d)
p1 = PCA()(data)
p2 = PCA()(data)
# copy projection
p2.domain[0].compute_value.compute_shared.projection = \
p1.domain[0].compute_value.compute_shared.projection
p2.proj = p1.proj
# reset hash caches because object were hacked
# pylint: disable=protected-access
p1.domain._hash = None
p2.domain._hash = None
p1.domain[0].compute_value.compute_shared._hash = None
p2.domain[0].compute_value.compute_shared._hash = None
self.assertEqual(p1, p2)
self.assertEqual(p1.domain, p2.domain)
self.assertEqual(hash(p1), hash(p2))
self.assertEqual(hash(p1.domain), hash(p2.domain))
|
