質問編集履歴
2
コード
test
CHANGED
|
File without changes
|
test
CHANGED
|
@@ -10,11 +10,7 @@
|
|
|
10
10
|
|
|
11
11
|
print "Finished Importing Numpy"
|
|
12
12
|
|
|
13
|
-
#from scipy import signal
|
|
14
13
|
|
|
15
|
-
#import scipy as sp
|
|
16
|
-
|
|
17
|
-
#print "Finished Importing Scipy"
|
|
18
14
|
|
|
19
15
|
import re #regular expression
|
|
20
16
|
|
|
@@ -28,11 +24,7 @@
|
|
|
28
24
|
|
|
29
25
|
print "Finished Importing Counter"
|
|
30
26
|
|
|
31
|
-
#import matplotlib.pyplot as plt
|
|
32
|
-
|
|
33
27
|
import time
|
|
34
|
-
|
|
35
|
-
#import math
|
|
36
28
|
|
|
37
29
|
|
|
38
30
|
|
|
@@ -55,8 +47,6 @@
|
|
|
55
47
|
print "eigenvals"
|
|
56
48
|
|
|
57
49
|
la = np.linalg.eigvalsh(np.corrcoef(self.dataarr))#np.corrcoef(self.dataarr)#
|
|
58
|
-
|
|
59
|
-
#U, s, V = np.linalg.svd(np.corrcoef(self.dataarr))#, full_matrices=False)
|
|
60
50
|
|
|
61
51
|
return la
|
|
62
52
|
|
|
@@ -84,16 +74,4 @@
|
|
|
84
74
|
|
|
85
75
|
laave = laave + s
|
|
86
76
|
|
|
87
|
-
#print s
|
|
88
|
-
|
|
89
77
|
print laave/100
|
|
90
|
-
|
|
91
|
-
#print np.corrcoef(np.random.randn(1600,100))
|
|
92
|
-
|
|
93
|
-
#sum=0.
|
|
94
|
-
|
|
95
|
-
#for num in range(0, 1000000000):
|
|
96
|
-
|
|
97
|
-
# sum=sum+np.random.randn()
|
|
98
|
-
|
|
99
|
-
#print sum/1000000000
|
1
コード追加
test
CHANGED
|
File without changes
|
test
CHANGED
|
@@ -5,3 +5,95 @@
|
|
|
5
5
|
の演算で、巨大な行列の固有値を求めるところでした。
|
|
6
6
|
|
|
7
7
|
何か解決策があれば教えてください。
|
|
8
|
+
|
|
9
|
+
import numpy as np
|
|
10
|
+
|
|
11
|
+
print "Finished Importing Numpy"
|
|
12
|
+
|
|
13
|
+
#from scipy import signal
|
|
14
|
+
|
|
15
|
+
#import scipy as sp
|
|
16
|
+
|
|
17
|
+
#print "Finished Importing Scipy"
|
|
18
|
+
|
|
19
|
+
import re #regular expression
|
|
20
|
+
|
|
21
|
+
print "Finished Importing Regex"
|
|
22
|
+
|
|
23
|
+
import struct
|
|
24
|
+
|
|
25
|
+
print "Finished Importing Struct"
|
|
26
|
+
|
|
27
|
+
from collections import Counter
|
|
28
|
+
|
|
29
|
+
print "Finished Importing Counter"
|
|
30
|
+
|
|
31
|
+
#import matplotlib.pyplot as plt
|
|
32
|
+
|
|
33
|
+
import time
|
|
34
|
+
|
|
35
|
+
#import math
|
|
36
|
+
|
|
37
|
+
|
|
38
|
+
|
|
39
|
+
class Factoranalysis:
|
|
40
|
+
|
|
41
|
+
def __init__(self,dataarr):
|
|
42
|
+
|
|
43
|
+
print "StartinitFA"
|
|
44
|
+
|
|
45
|
+
self.dataarr=dataarr
|
|
46
|
+
|
|
47
|
+
self.numval=dataarr.shape[0]
|
|
48
|
+
|
|
49
|
+
self.numobs=dataarr.shape[1]
|
|
50
|
+
|
|
51
|
+
print "FAinitFIN"
|
|
52
|
+
|
|
53
|
+
def eigenvals(self):
|
|
54
|
+
|
|
55
|
+
print "eigenvals"
|
|
56
|
+
|
|
57
|
+
la = np.linalg.eigvalsh(np.corrcoef(self.dataarr))#np.corrcoef(self.dataarr)#
|
|
58
|
+
|
|
59
|
+
#U, s, V = np.linalg.svd(np.corrcoef(self.dataarr))#, full_matrices=False)
|
|
60
|
+
|
|
61
|
+
return la
|
|
62
|
+
|
|
63
|
+
def parallelanalysis(self,repnum):
|
|
64
|
+
|
|
65
|
+
laave=0.
|
|
66
|
+
|
|
67
|
+
for num in range(0,repnum):
|
|
68
|
+
|
|
69
|
+
print laave
|
|
70
|
+
|
|
71
|
+
laave = np.linalg.eigvalsh(np.corrcoef(np.random.randn(self.numval,self.numval))) + laave
|
|
72
|
+
|
|
73
|
+
return laave/repnum
|
|
74
|
+
|
|
75
|
+
|
|
76
|
+
|
|
77
|
+
if __name__ == '__main__':
|
|
78
|
+
|
|
79
|
+
laave=0.
|
|
80
|
+
|
|
81
|
+
for num in range(0,100):
|
|
82
|
+
|
|
83
|
+
U, s, V = np.linalg.svd(np.corrcoef(np.random.randn(100,100)))
|
|
84
|
+
|
|
85
|
+
laave = laave + s
|
|
86
|
+
|
|
87
|
+
#print s
|
|
88
|
+
|
|
89
|
+
print laave/100
|
|
90
|
+
|
|
91
|
+
#print np.corrcoef(np.random.randn(1600,100))
|
|
92
|
+
|
|
93
|
+
#sum=0.
|
|
94
|
+
|
|
95
|
+
#for num in range(0, 1000000000):
|
|
96
|
+
|
|
97
|
+
# sum=sum+np.random.randn()
|
|
98
|
+
|
|
99
|
+
#print sum/1000000000
|