/
sim.py
executable file
·150 lines (116 loc) · 2.84 KB
/
sim.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
#!/usr/bin/python
import random
import numpy as np
import math
NUM_SAMPLES = 500;
DIM_SAMPLES = 11;
eta = 1
def trainOverData(data, labels, w, wList):
errorCount = 0
for i, x in enumerate(data):
if np.dot(w, x)*labels[i] <= 0:
w = w + eta*labels[i]*x
errorCount += 1
wList.append(w)
return w, errorCount
def testOverData(data, labels, w):
errorCount = 0
for i, x in enumerate(data):
if np.dot(w, x)*labels[i] <= 0:
errorCount += 1
return errorCount
def randomSample(length):
return [int(random.getrandbits(1)*2-1) for i in xrange(length)]
def labelList(data):
labels = []
for x in data:
labels.append(x[0])
return labels
def labelList2(data):
labels = []
for x in data:
label = 1 if sum(x) > 0 else -1
labels.append(label)
return labels
def labelList3(data):
labels = []
for x in data:
r = random.randint(-3, 3)
label = 1 if sum(x)+r > 0 else -1
labels.append(label)
return labels
def testOverDataVoting(data, labels, wList):
errorCount = 0
for i, x in enumerate(data):
vote = 0
for w in wList:
vote += math.copysign(1, np.dot(w, x))
if vote*labels[i] <= 0:
errorCount += 1
return errorCount
def avgW(wList):
avg = np.zeros(11)
for x in wList:
avg += x/1000.0
return avg
def partCTest(wList):
w_1000 = wList[-1]
w_avg = avgW(wList)
train_data = [np.array(randomSample(DIM_SAMPLES)) for i in xrange(NUM_SAMPLES)]
labels = labelList3(train_data)
accuracies = []
accuracies.append(testOverData(train_data, labels, w_1000))
accuracies.append(testOverData(train_data, labels, w_avg))
accuracies.append(testOverDataVoting(train_data, labels, wList))
return accuracies
def sim(labelType):
# Init data
data = [np.array(randomSample(DIM_SAMPLES)) for i in xrange(NUM_SAMPLES)]
# Create labels
if labelType == 2:
labels = labelList2(data)
else:
labels = labelList(data)
# Init w
w = np.array([0 for i in xrange(DIM_SAMPLES)])
wList = []
cumError = 0
errorCount = int(True)
epochCount = 0
while errorCount:
w, errorCount = trainOverData(data, labels, w, wList)
cumError += errorCount
# print 'w: ', str(w)
epochCount += 1
print 'In %d epochs with %d errors' % (epochCount, cumError)
return
def sim3():
# Init data
data = [np.array(randomSample(DIM_SAMPLES)) for i in xrange(NUM_SAMPLES)]
# Create labels
labels = labelList3(data)
# Init w
w = np.array([0 for i in xrange(DIM_SAMPLES)])
wList = []
errorCount = int(True)
epochCount = 0
while errorCount:
w, errorCount = trainOverData(data, labels, w, wList)
# print 'w: ', str(w)
# print '%d errors' % errorCount
epochCount += 1
if epochCount >= 2: errorCount = 0
# print 'In %d epochs' % epochCount
return partCTest(wList)
def run3():
results = []
for i in xrange(10):
results.append(sim3())
avg = map(np.mean, zip(*results))
print avg
def main():
sim(1)
sim(2)
run3()
if __name__ == '__main__':
main()