-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcalculate_glass_binary_bayesian.py
378 lines (338 loc) · 14.6 KB
/
calculate_glass_binary_bayesian.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
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
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
import numpy as np
#import pandas as pd
from numpy.linalg import *
from math import isnan, isinf
from sklearn.metrics import confusion_matrix
from sklearn.metrics import accuracy_score
#import math
def calculate_accuracy(data,features,test,totalClass,foldSize):
#mean calculation
#class
W1 = data[data[:, features] == totalClass[0]]
W1 = W1[:, :features]
#print "class-1 dimension", W1.shape
mean_class1 = mean_calulate(W1)
print ("Class-1 mean: ", mean_class1.shape)
W1_cov = find_cov(W1,mean_class1,features)
print ("COV class-1: ", W1_cov.shape)
#print np.isnan(W1_cov)
W1_cov_inv = pinv(W1_cov)
print ("Inverse COV class-1: ", W1_cov_inv.shape)
W2 = data[data[:, features] == totalClass[1]]
W2 = W2[:, :features]
print ("class-2 dimension", W2.shape)
mean_class2 = mean_calulate(W2)
print ("Class-2 mean: ", mean_class2.shape)
W2_cov = find_cov(W2,mean_class2,features)
print ("COV class-2: ", W2_cov.shape)
W2_cov_inv = pinv(W2_cov)
print ("Inverse COV class-2: ", W2_cov_inv.shape)
W3 = data[data[:, features] == totalClass[2]]
W3 = W3[:, :features]
#print "class-3 dimension", W3.shape
mean_class3 = mean_calulate(W3)
print ("Class-3 mean: ", mean_class3.shape)
W3_cov = find_cov(W3,mean_class3,features)
print ("COV class-3: ", W3_cov.shape)
W3_cov_inv = pinv(W3_cov)
print ("Inverse COV class-3: ", W3_cov_inv.shape)
W4 = data[data[:, features] == totalClass[3]]
W4 = W4[:, :features]
#print "class-4 dimension", W4.shape
mean_class4 = mean_calulate(W4)
print ("Class-4 mean: ", mean_class4.shape)
W4_cov = find_cov(W4,mean_class4,features)
print ("COV class-4: ", W4_cov.shape)
W4_cov_inv = pinv(W4_cov)
print ("Inverse COV class-4: ", W4_cov_inv.shape)
'''
W5 = data[data[:, features] == totalClass[4]]
W5 = W5[:, :features]
#print "class-5 dimension", W5.shape
mean_class5 = mean_calulate(W5)
print "Class-5 mean: ", mean_class5.shape
W5_cov = find_cov(W5,mean_class5,features)
print "COV class-5: ", W5_cov.shape
W5_cov_inv = pinv(W5_cov)
print "Inverse COV class-5: ", W5_cov_inv.shape
W6 = data[data[:, features] == totalClass[5]]
W6 = W6[:, :features]
#print "class-6 dimension", W6.shape
mean_class6 = mean_calulate(W6)
print "Class-6 mean: ", mean_class6.shape
W6_cov = find_cov(W6,mean_class6,features)
print "COV class-6: ", W6_cov.shape
W6_cov_inv = pinv(W6_cov)
print "Inverse COV class-6: ", W6_cov_inv.shape
'''
classSeparated = []
#print test.shape
test_X = np.transpose(test[:,:features])
print (test_X.shape)
test_Y = np.reshape(test[:, features],((foldSize),1))
for i in range(np.size(test_X,1)):
#compare class 1 and 2
classifier1 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W1_cov_inv,W2_cov_inv,W1_cov,W2_cov,mean_class1,mean_class2)
#print "classifier -1 ", classifier1
if classifier1 > 0:
#compare class 1 and 3
classifier2 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W1_cov_inv,W3_cov_inv,W1_cov,W3_cov,mean_class1,mean_class3)
#print "classifier - 2", classifier2
if classifier2 > 0:
#compare class 1 and 4
classifier3 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W1_cov_inv,W4_cov_inv,W1_cov,W4_cov,mean_class1,mean_class4)
if classifier3 > 0:
classSeparated.append(totalClass[0])
else:
classSeparated.append(totalClass[3])
else:
classifier3 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W3_cov_inv,W4_cov_inv,W3_cov,W4_cov,mean_class3,mean_class4)
if classifier3 > 0:
classSeparated.append(totalClass[2])
else:
classSeparated.append(totalClass[3])
else:
#compare class 2 and 3
classifier2 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W2_cov_inv,W3_cov_inv,W2_cov,W3_cov,mean_class2,mean_class3)
#print "classifier - 3", classifier3
if classifier2 > 0:
#compare class 2 and 4
classifier3 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W2_cov_inv,W4_cov_inv,W2_cov,W4_cov,mean_class2,mean_class4)
if classifier3 > 0:
classSeparated.append(totalClass[1])
else:
classSeparated.append(totalClass[3])
else:
classifier3 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W3_cov_inv,W4_cov_inv,W3_cov,W4_cov,mean_class3,mean_class4)
if classifier3 > 0:
classSeparated.append(totalClass[2])
else:
classSeparated.append(totalClass[3])
classSeparated = np.array(classSeparated)
#print classSeparated, test_Y
print ("The Confusion Matrix: ")
print (confusion_matrix(test_Y, classSeparated))
print ("The Accuracy Score: ", accuracy_score(test_Y, classSeparated)*100)
return find_accuracy(test_Y,classSeparated)
def calculate_accuracy_naive(data,features,test,totalClass,foldSize):
#mean calculation
#class
W1 = data[data[:, features] == totalClass[0]]
W1 = W1[:, :features]
#print "class-1 dimension", W1.shape
mean_class1 = mean_calulate(W1)
print ("Class-1 mean: ", mean_class1.shape)
W1_cov = find_cov_naive(W1,mean_class1,features)
print ("COV naive class-1: ", W1_cov.shape)
#print np.isnan(W1_cov)
W1_cov_inv = pinv(W1_cov)
print ("Inverse COV class-1: ", W1_cov_inv.shape)
W2 = data[data[:, features] == totalClass[1]]
W2 = W2[:, :features]
#print "class-2 dimension", W2.shape
mean_class2 = mean_calulate(W2)
print ("Class-2 mean: ", mean_class2.shape)
W2_cov = find_cov_naive(W2,mean_class2,features)
print ("COV class-2: ", W2_cov.shape)
W2_cov_inv = pinv(W2_cov)
print ("Inverse COV class-2: ", W2_cov_inv.shape)
W3 = data[data[:, features] == totalClass[2]]
W3 = W3[:, :features]
#print "class-3 dimension", W3.shape
mean_class3 = mean_calulate(W3)
print ("Class-3 mean: ", mean_class3.shape)
W3_cov = find_cov_naive(W3,mean_class3,features)
print ("COV class-3: ", W3_cov.shape)
W3_cov_inv = pinv(W3_cov)
print ("Inverse COV class-3: ", W3_cov_inv.shape)
W4 = data[data[:, features] == totalClass[3]]
W4 = W4[:, :features]
#print "class-4 dimension", W4.shape
mean_class4 = mean_calulate(W4)
print ("Class-4 mean: ", mean_class4.shape)
W4_cov = find_cov(W4,mean_class4,features)
print ("COV class-4: ", W4_cov.shape)
W4_cov_inv = pinv(W4_cov)
print ("Inverse COV class-4: ", W4_cov_inv.shape)
'''
W5 = data[data[:, features] == totalClass[4]]
W5 = W5[:, :features]
#print "class-5 dimension", W5.shape
mean_class5 = mean_calulate(W5)
print "Class-5 mean: ", mean_class5.shape
W5_cov = find_cov(W5,mean_class5,features)
print "COV class-5: ", W5_cov.shape
W5_cov_inv = pinv(W5_cov)
print "Inverse COV class-5: ", W5_cov_inv.shape
W6 = data[data[:, features] == totalClass[5]]
W6 = W6[:, :features]
#print "class-6 dimension", W6.shape
mean_class6 = mean_calulate(W6)
print "Class-6 mean: ", mean_class6.shape
W6_cov = find_cov(W6,mean_class6,features)
print "COV class-6: ", W6_cov.shape
W6_cov_inv = pinv(W6_cov)
print "Inverse COV class-6: ", W6_cov_inv.shape
'''
classSeparated = []
test_X = np.transpose(test[:,:9])
#print test_X.shape
test_Y = np.reshape(test[:, 9],((foldSize),1))
#print np.size(test_X,1)
for i in range(np.size(test_X,1)):
#compare class 1 and 2
classifier1 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W1_cov_inv,W2_cov_inv,W1_cov,W2_cov,mean_class1,mean_class2)
#print "classifier -1 ", classifier1
if classifier1 > 0:
#compare class 1 and 3
classifier2 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W1_cov_inv,W3_cov_inv,W1_cov,W3_cov,mean_class1,mean_class3)
#print "classifier - 2", classifier2
if classifier2 > 0:
#compare class 1 and 4
classifier3 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W1_cov_inv,W4_cov_inv,W1_cov,W4_cov,mean_class1,mean_class4)
if classifier3 > 0:
classSeparated.append(totalClass[0])
else:
classSeparated.append(totalClass[3])
else:
classifier3 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W3_cov_inv,W4_cov_inv,W3_cov,W4_cov,mean_class3,mean_class4)
if classifier3 > 0:
classSeparated.append(totalClass[2])
else:
classSeparated.append(totalClass[3])
else:
#compare class 2 and 3
classifier2 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W2_cov_inv,W3_cov_inv,W2_cov,W3_cov,mean_class2,mean_class3)
#print "classifier - 3", classifier3
if classifier2 > 0:
#compare class 2 and 4
classifier3 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W2_cov_inv,W4_cov_inv,W2_cov,W4_cov,mean_class2,mean_class4)
if classifier3 > 0:
classSeparated.append(totalClass[1])
else:
classSeparated.append(totalClass[3])
else:
classifier3 = calculateClassifier(np.reshape(test_X[:,i],(9,1)),W3_cov_inv,W4_cov_inv,W3_cov,W4_cov,mean_class3,mean_class4)
if classifier3 > 0:
classSeparated.append(totalClass[2])
else:
classSeparated.append(totalClass[3])
classSeparated = np.array(classSeparated)
#print classSeparated, test_Y
print ("The Confusion Matrix: ")
print (confusion_matrix(test_Y, classSeparated))
print ("The Accuracy Score: ", accuracy_score(test_Y, classSeparated)*100)
return find_accuracy(test_Y,classSeparated)
def mean_calulate(X):
N = np.size(X,0) #0: total row
M = np.size(X,1) #1: total col
#print X
summation = np.sum(X, axis=0) #0: row wise sum: np.sum([[0, 1], [0, 5]], axis=0) = array([0, 6])
#print summation, N
mean = summation / N
return np.reshape(mean,(M,1))
def find_cov(W1_X, W1_mean,features):
W1_X_transpose = np.transpose(W1_X) #make it 4*n
W1_cov = np.empty([features,features])
#print "Shape of class-1 transpose matrix: ", W1_X_transpose.shape
test = []
#print W1_X_transpose
#print W1_X_transpose.shape
W1_X_mean_subtract = np.subtract(W1_X_transpose, W1_mean)
print ("class-1 after calculate subtract: ", W1_X_mean_subtract.shape)
#print W1_X_mean_subtract
elements = np.size(W1_X_mean_subtract,0)
#print elements
for i in range(np.size(W1_X_mean_subtract,1)): #for features
select1 = np.reshape(W1_X_mean_subtract[:,i],(elements,1))
#print select1.shape
select2 = select1.dot(np.transpose(select1))
#print select2.shape
test.append(np.array(select2))
return mean_calulate2(W1_cov,test)
def find_cov_naive(W1_X, W1_mean,features):
W1_X_transpose = np.transpose(W1_X) #make it 4*n
W1_cov = np.empty([features,features])
#print "Shape of class-1 transpose matrix: ", W1_X_transpose.shape
test = []
#print W1_X_transpose
#print W1_X_transpose.shape
W1_X_mean_subtract = np.subtract(W1_X_transpose, W1_mean)
print ("class-1 after calculate subtract: ", W1_X_mean_subtract.shape)
#print W1_X_mean_subtract
elements = np.size(W1_X_mean_subtract,0)
#print elements
for i in range(np.size(W1_X_mean_subtract,1)): #for features
select1 = np.reshape(W1_X_mean_subtract[:,i],(elements,1))
#print select1.shape
select2 = select1.dot(np.transpose(select1))
#print select2.shape
test.append(np.array(select2))
W1_cov = mean_calulate2(W1_cov,test)
#I = np.identity(4)
#print W1_cov
#print I
W1_cov = np.diag(np.diag(W1_cov))
#print "xxxx:", W1_cov
return W1_cov
def mean_calulate2(X,list):
N = len(list) #0: total row
#print N
summation = np.empty(X.shape)
#print summation.shape
for i in range(N):
summation = summation + list[i]
#print summation
cov = summation / (N-1)
cov = cov.astype(np.float64)
#fix for numpy.linalg.LinAlgError: SVD did not converge (nan)
cov = np.nan_to_num(cov)
return np.array(cov)
def calculateClassifier(X,covInv1,covInv2,cov1,cov2,mean1,mean2):
#print "classifier"
A = (covInv2-covInv1)
#print A.shape, X.shape, covInv1.shape, cov1.shape, mean1.shape
A = (np.transpose(X).dot(A)).dot(X)
#print firstpart.shape
B = ((np.transpose(mean2).dot(covInv2)) - (np.transpose(mean1).dot(covInv1)))
#print B.shape
B = -2 * (B.dot(X))
if isnan(A) or isinf(A): A = 0
if isnan(B) or isinf(B): B = 0
X2 = find_pseudodeterminant(cov2)
#print X2
variableLog2 = 0.0
if (X2>0.0 and (isinstance(X2, complex) != True)):
if isnan(np.log(X2)) or isinf(np.log(X2)):
variableLog2 = 0.0
else:
variableLog2 = np.log(X2)
X1 = find_pseudodeterminant(cov1)
#print X1
variableLog1 = 0.0
if (X1>0.0 and (isinstance(X1, complex) != True)):
if isnan(np.log(X1)) or isinf(np.log(X1)) or isinstance(X1, complex):
variableLog1 = 0.0
else:
variableLog1 = np.log(X1)
variableLog = variableLog2 - variableLog1
C = variableLog + ((np.transpose(mean2).dot(covInv2)).dot(mean2)) - ((np.transpose(mean1).dot(covInv1)).dot(mean1))
#print C
if isnan(A+B+C):
return 0.0
else:
return int(round(A+B+C))
def find_accuracy(test_Y,test_pred):
result = (test_Y == test_pred)
trueDetection = (float)(np.sum(result))
size = (int)(np.size(test_pred,0))
#print trueDetection, size, trueDetection/size
accuracy = (trueDetection/size)*100
return accuracy
def find_pseudodeterminant(covarianceMatrix):
#First compute the eigenvalues of your matrix
eig_values = np.linalg.eig(covarianceMatrix)
#Then compute the product of the non-zero eigenvalues (this equals the pseudo-determinant value of the matrix)
pseudo_determinent = np.product(eig_values[eig_values > 1e-12])
#print pseudo_determinent
return pseudo_determinent