-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtraining.py
150 lines (117 loc) · 4.28 KB
/
training.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
import time
import math
import numpy as np
import sys
import matplotlib
import Pycluster
import matplotlib.pyplot as plt
from scipy import stats
from ghmm import *
def matrices(state_num, output_num):
sigma = IntegerRange(0, output_num)
# matrica A #####################################################################
A = np.matrix(np.identity(state_num)) / float(2)
for i in range(state_num - 1):
A[i, i+1] = float(1)/2
A[state_num-1,state_num-1] = 1
#print A
# matrica B #####################################################################
B = np.ones((state_num, output_num)) / float(output_num)
#print B
# pi ############################################################################
A = A.tolist()
B = B.tolist()
pi = np.ones((state_num))/float(state_num)
pi = pi.tolist()
#pi[0] = 1
#print pi
return sigma, A, B, pi
def sort_labels(labels):
existing_labels = [labels[0]]
for label in labels:
try:
existing_labels.index(label)
except ValueError:
existing_labels.append(label)
sorted_labels = sorted(existing_labels)
ordered_labels = np.asarray([sorted_labels[existing_labels.index(element)] for element in labels])
return ordered_labels
def cluster(fname, nclust):
fh = open(fname, 'r')
lines = fh.readlines()
fh.close()
clusters = int(nclust)
points = []
points_r = []
dates = []
volumes = []
close_prices = []
for i in range(len(lines)):
if i <= 1:
continue
line_c = lines[i-1].strip().split(',')
close_price = float(line_c[0])
volume = float(line_c[1])
points_r.append((close_price, volume))
volumes.append(volume)
close_prices.append(close_price)
#dates.append(line_c[0])
volume_z= np.array(volumes)
#volume_z = stats.zscore(a)
close_price_z = np.array(close_prices)
#close_price_z = stats.zscore(a)
points = zip(close_price_z, volume_z)
init_data = []
k = len(points) / (nclust)
for i in range(nclust - 1):
for j in range(k):
init_data.append(i)
while(len(points) != len(init_data)):
init_data.append(nclust-1)
#print(clusters)
labels, error, nfound = Pycluster.kcluster(points, clusters, None, None, 0, 1, 'a', 'e', init_data)
labels_sorted = sort_labels(labels)
#print('Labels: ')
print labels_sorted
return labels_sorted
if __name__ == '__main__':
state_num = input('Broj stanja: ')
output_num = input('Broj izlaza: ')
dataset_num = 101 # broj fileova data seta
clust_data = []
clust_set = []
gesture_seq = []
gesture_treshold = 0
for i in range(dataset_num):
clust_data.append(cluster("./geste/Drink/gest" + str(i) + ".txt", output_num)) # file-ovi za trening
#clust_data.append(cluster("gest" + str(i) + ".txt", output_num))
for i in range(dataset_num):
clust_set = np.concatenate((clust_set, clust_data[i]), 0)
[sigma, A, B, pi] = matrices(state_num, output_num)
m = HMMFromMatrices(sigma, DiscreteDistribution(sigma), A, B, pi)
#print m
# train sequence #################################################################
train_seq = EmissionSequence(sigma, clust_set.tolist())
#print train_seq
m.baumWelch(train_seq)
# m.write('./geste/Drink/m_file.xml') # izrada m filea
#print m
for i in range(dataset_num):
gesture_seq.append(EmissionSequence(sigma, clust_data[i].tolist()))
#for i in range(dataset_num):
# print m.loglikelihood(gesture_seq[i])
# gesture_treshold = gesture_treshold + m.loglikelihood(gesture_seq[i])
#gesture_treshold = gesture_treshold / float(dataset_num)
for i in range(dataset_num):
#print('###########################################')
#print m.viterbi(gesture_seq[i])
gesture_treshold = gesture_treshold + m.viterbi(gesture_seq[i])[1]
gesture_treshold = gesture_treshold / float(dataset_num)
#print gesture_treshold
# f = open('./geste/Drink/gesture_file.txt', 'w') # podaci za gestu, treshold i broj stanja
# f.writelines(str(state_num))
# f.write('\n')
# f.write(str(output_num))
# f.write('\n')
# f.write(str(gesture_treshold))
# f.close()