-
Notifications
You must be signed in to change notification settings - Fork 3
/
test_spectral_lda.py
71 lines (54 loc) · 1.88 KB
/
test_spectral_lda.py
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
''' Test Spectral LDA '''
import numpy as np
from scipy.sparse import csr_matrix
from spectral_lda import spectral_lda
from test_cumulants import simulate_word_count_vectors
def test_spectral_lda():
''' Simple test cases '''
gen_alpha = [10, 5, 2]
gen_k = len(gen_alpha)
vocab_size = 50
n_docs = 5000
gen_beta = np.random.rand(vocab_size, gen_k)
gen_beta /= gen_beta.sum(axis=0)
docs = simulate_word_count_vectors(gen_alpha, gen_beta, n_docs, 500, 1000)
for n_partitions in [1, 3]:
k = gen_k
alpha0 = np.sum(gen_alpha[:k])
alpha, beta = spectral_lda(docs, alpha0, k,
n_partitions=n_partitions)
print('Generative alpha:')
print(gen_alpha)
print('Fitted alpha:')
print(alpha)
print('Generative beta:')
print(gen_beta)
print('Fitted beta:')
print(beta)
assert np.all(np.linalg.norm(gen_beta[:, :k] - beta, axis=0) < 0.2)
def test_spectral_lda_csr_matrix():
''' Simple test cases '''
gen_alpha = [10, 5, 2]
gen_k = len(gen_alpha)
vocab_size = 50
n_docs = 5000
gen_beta = np.random.rand(vocab_size, gen_k)
for j in range(gen_k):
gen_beta[(j * 5):((j + 1) * 5), j] += 1
gen_beta /= gen_beta.sum(axis=0)
docs = simulate_word_count_vectors(gen_alpha, gen_beta, n_docs, 500, 1000)
docs = csr_matrix(docs)
for n_partitions in [1, 3]:
k = gen_k
alpha0 = np.sum(gen_alpha[:k])
alpha, beta = spectral_lda(docs, alpha0, k,
n_partitions=n_partitions)
print('Generative alpha:')
print(gen_alpha)
print('Fitted alpha:')
print(alpha)
print('Generative beta:')
print(gen_beta)
print('Fitted beta:')
print(beta)
assert np.all(np.linalg.norm(gen_beta[:, :k] - beta, axis=0) < 0.2)