-
Notifications
You must be signed in to change notification settings - Fork 3
/
readdata.py
222 lines (201 loc) · 8.71 KB
/
readdata.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
# -*- coding: utf-8 -*-
"""
Created on Fri Jun 24 11:27:54 2016
@author: Stephen-Lu
"""
import os
import numpy as np
import matplotlib.pyplot as plt
from keras.layers import Input, Dense, Convolution2D, MaxPooling2D, UpSampling2D
from keras.layers.local import LocallyConnected2D
from keras.models import Model
from utils import tile_raster_images
import keras
from sklearn import svm
from sklearn import cross_validation
def load(dir, rescale = False):
direc = dir+'/train/'
files = os.listdir(direc)
channel_num_train = len(files)
datum_train = []
for f in files:
data = open(direc + f, 'r').readlines()
data = [np.asarray(map(float, each.strip().split('\t'))) for each in data]
if f == files[0]:
label_train = [each[0] for each in data]
l = [each.size for each in data]
else:
assert label_train == [each[0] for each in data]
l == [each.size for each in data]
data = [(each[1:]-np.mean(each[1:]))/(np.std(each[1:])+0.001) for each in data]
if rescale:
data = [(each[1:]-np.min(each[1:]))/(np.max(each[1:]) - np.min(each[1:] + +0.001)) for each in data]
datum_train.append(data)
# del data
#Reshape the data
sample_num_train = len(datum_train[0])
length_train = max(l)
direc = dir+'/test/'
files = os.listdir(direc)
channel_num_test = len(files)
datum_test = []
for f in files:
data = open(direc + f, 'r').readlines()
data = [np.asarray(map(float, each.strip().split('\t'))) for each in data]
if f == files[0]:
label_test = [each[0] for each in data]
l = [each.size for each in data]
else:
assert label_test == [each[0] for each in data]
l == [each.size for each in data]
data = [(each[1:]-np.mean(each[1:]))/(np.std(each[1:])+0.001) for each in data]
if rescale:
data = [(each[1:]-np.min(each[1:]))/(np.max(each[1:]) - np.min(each[1:] + +0.001)) for each in data]
datum_test.append(data)
# del data
#Reshape the data
sample_num_test = len(datum_test[0])
length_test = max(l)
assert channel_num_test == channel_num_train
channel_num = channel_num_test
length = max(length_train, length_test)
def form_data(datum, sample_num, label):
mat = []
for s in xrange(sample_num):
vec = np.asarray([])
for c in xrange(channel_num):
vec = np.hstack((vec,np.hstack((datum[c][s],np.full(length - datum[c][s].size,datum[c][s].min())))))
mat.append(vec)
mat = np.asarray(mat, dtype = np.float32)
label = np.asarray(label, dtype = np.float32)
return mat, label
x_train, y_train = form_data(datum_train, sample_num_train, label_train)
x_test, y_test = form_data(datum_test, sample_num_test, label_test)
return (channel_num, length), (x_train, y_train), (x_test, y_test)
names = ['AUSLAN_MTS', 'JP_MTS','Libra_MTS','LP_MTS','MOCAP_MTS', 'wafer_MTS']
names = ['wafer_MTS']
for name in names:
#Read all data in memory and do normalization
direc = 'MTSdata/'+ name
shape, (x_train, y_train), (x_test, y_test) = load(direc, True)
#Convolutional AutoEncoder
input_img = Input(shape=(1, shape[0], shape[1]))
#x = Convolution2D(16, 1, 1, activation='relu', border_mode='same')(input_img)
#x = MaxPooling2D((2, 1), border_mode='same')(x)
#x = Convolution2D(8, 1, 1, activation='relu', border_mode='same')(x)
#x = MaxPooling2D((2, 1), border_mode='same')(x)
#x = Convolution2D(8, 1, 1, activation='relu', border_mode='same')(x)
#encoded = MaxPooling2D((2, 1), border_mode='same')(x)
#
## at this point the representation is (8, 4, 4) i.e. 128-dimensional
#
#x = Convolution2D(8, 1, 1, activation='relu', border_mode='same')(encoded)
#x = UpSampling2D((2, 1))(x)
#x = Convolution2D(8, 1, 1, activation='relu', border_mode='same')(x)
#x = UpSampling2D((2, 1))(x)
#x = Convolution2D(16, 1, 1, activation='relu')(x)
#x = UpSampling2D((2, 1))(x)
#decoded = Convolution2D(1, 3, 3, activation='sigmoid', border_mode='same')(x)
#x = Convolution2D(16, shape[0], 6, activation='relu', border_mode='same')(input_img)
#encoded = MaxPooling2D((shape[0], 1), border_mode='valid')(x)
#x = Convolution2D(16,shape[0], 6, activation='relu', border_mode='same')(encoded)
#x = UpSampling2D((shape[0], 1))(x)
#decoded = Convolution2D(1, 1, 1, activation='sigmoid', border_mode='same')(x)
x = Convolution2D(16, 3, 3, activation='relu', border_mode='same')(input_img)
x = MaxPooling2D((shape[0]/2, 1), border_mode='valid')(x)
x = Convolution2D(8, 3, 3, activation='relu', border_mode='same')(input_img)
encoded = MaxPooling2D((shape[0]/2, 1), border_mode='valid')(x)
x = Convolution2D(8,3, 3, activation='relu', border_mode='same')(encoded)
x = UpSampling2D((shape[0]/2, 1))(x)
x = Convolution2D(16,3, 3, activation='relu', border_mode='same')(encoded)
x = UpSampling2D((shape[0]/2, 1))(x)
decoded = Convolution2D(1, 3, 3, activation='sigmoid', border_mode='same')(x)
autoencoder = Model(input_img, decoded)
optimizer = keras.optimizers.Adadelta(lr=.1, rho=0.95, epsilon=1e-08)
autoencoder.compile(optimizer=optimizer, loss='binary_crossentropy')
x_train = np.reshape(x_train, (len(x_train), 1, shape[0], shape[1]))
x_test = np.reshape(x_test, (len(x_test), 1, shape[0], shape[1]))
autoencoder.fit(x_train, x_train,
nb_epoch=200,
batch_size=5,
verbose=1,
validation_data=(x_test, x_test))
decoded_imgs = autoencoder.predict(x_test)
#
encoder = Model(input = input_img, output = encoded)
encoded_train = encoder.predict(x_train)
encoded_test = encoder.predict(x_test)
#%%
n = 5
plt.figure(figsize=(20, 4))
for i in range(n):
# display original
ax = plt.subplot(2, n, i + 1)
#plt.imshow(x_test[i].reshape(shape[0], shape[1]), extent=[0,shape[1],0,100])
plt.plot(np.arange(shape[1]),x_test[i].reshape(shape[0], shape[1])[0], np.arange(shape[1]), x_test[i].reshape(shape[0], shape[1])[1],
np.arange(shape[1]),x_test[i].reshape(shape[0], shape[1])[2], np.arange(shape[1]), x_test[i].reshape(shape[0], shape[1])[3],
np.arange(shape[1]),x_test[i].reshape(shape[0], shape[1])[4], np.arange(shape[1]), x_test[i].reshape(shape[0], shape[1])[5])
plt.gray()
ax.get_xaxis().set_visible(False)
#ax.get_yaxis().set_visible(False)
# display reconstruction
ax = plt.subplot(2, n, i + n +1)
#plt.imshow(decoded_imgs[i].reshape(shape[0], shape[1]), extent=[0,shape[1],0,100])
plt.plot(np.arange(shape[1]),decoded_imgs[i].reshape(shape[0], shape[1])[0], np.arange(shape[1]), decoded_imgs[i].reshape(shape[0], shape[1])[1],
np.arange(shape[1]),decoded_imgs[i].reshape(shape[0], shape[1])[2], np.arange(shape[1]), decoded_imgs[i].reshape(shape[0], shape[1])[3],
np.arange(shape[1]),decoded_imgs[i].reshape(shape[0], shape[1])[4], np.arange(shape[1]), decoded_imgs[i].reshape(shape[0], shape[1])[5])
plt.gray()
ax.get_xaxis().set_visible(False)
#ax.get_yaxis().set_visible(False)
plt.show()
#%%
#n = 10
#plt.figure(figsize=(20, 8))
#for i in range(n):
# ax = plt.subplot(1, n, i+1)
# plt.imshow(encoded_imgs[i].reshape(4, 4 * 8).T)
# plt.gray()
# ax.get_xaxis().set_visible(False)
# ax.get_yaxis().set_visible(False)
#plt.show()
#%%
feature_train = encoded_train.reshape(encoded_train.shape[0],-1)
feature_test = encoded_test.reshape(encoded_test.shape[0],-1)
np.savez('draw_feature_train_'+ name,x = feature_train, y=y_train)
np.savez('draw_feature_test_'+ name ,x=feature_test, y=y_test)
clf = svm.LinearSVC()
clf.fit(feature_train, y_train)
print np.mean(cross_validation.cross_val_score(clf, feature_train, y_train, cv=len(y_train)*2/3))
print clf.score(feature_train, y_train)
print clf.score(feature_test, y_test)
##%%
#ff = 'D:\Dropbox\CoZzu\Conv_SAX\MTSdata\ECG_MTS\ECG_TRAIN'
#dd = open(ff, 'r').readlines()
#sample = np.NaN
#ft0 = open('ECG_TRAIN0', 'w')
#ft1 = open('ECG_TRAIN1', 'w')
#for each in dd:
# d = each.strip().split()
# if d[0] != sample:
# ft0.write('\n'+d[2]+'\t')
# ft1.write('\n'+d[2]+'\t')
# sample = d[0]
# ft0.write(d[3]+'\t')
# ft1.write(d[4]+'\t')
#ft0.close()
#ft1.close()
#%%
#n = 5
#plt.figure(figsize=(20, 4))
#for i in range(n):
# # display original
# ax = plt.subplot(2, n, i + 1)
# #plt.imshow(x_test[i].reshape(shape[0], shape[1]), extent=[0,shape[1],0,100])
# plt.plot(np.arange(199),dd[0].reshape(-1, 199)[i+3])
# ax.get_xaxis().set_visible(False)
# #ax.get_yaxis().set_visible(False)
# ax = plt.subplot(2, n, i + n + 1)
# #plt.imshow(x_test[i].reshape(shape[0], shape[1]), extent=[0,shape[1],0,100])
# plt.plot(np.arange(153),ccc[0].reshape(-1, 153)[i+5])
# ax.get_xaxis().set_visible(False)
#plt.show()