pattern-recognition/test.py at master · ngiachou/pattern-recognition · GitHub

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
from perceptron import *
from reading_file import *
from trainers import *

def tester(p, data):
	last = len(data[0])-1
	true_positive = 0
	false_positive = 0
	true_negative = 0
	false_negative = 0
	for v in data:
		#print v
		#raw_input()
		if v[last] == 1:
			if v[last] == p.fire(v[:last]):
				true_positive += 1
			else:
				print v
				false_negative += 1
		else:
			if v[last] == p.fire(v[:last]):
				true_negative += 1
			else:
				print v
				false_positive += 1
	print "Test  results"
	print "true  positive: "+str(true_positive)
	print "true  negative: "+str(true_negative)
	print "false positive: "+str(false_positive)
	print "false negative: "+str(false_negative)
	print "Summation     : "+str(true_positive+true_negative+false_positive+false_negative)

if __name__ == '__main__':
	p = Perceptron([1,1,1,1])
	print "Testing perceptron with weight vector: "+str(p.getWeightVector())
	print "Result for input vector (255,100,30,1): "+str(p.fire([255,100,30,1]))
	print "Result for input vector (200,110,50,1): "+str(p.fire([200,110,50,1]))
	print "Result for input vector (30,100,200,1): "+str(p.fire([30,100,200,1]))
	print "Result for input vector (120,235,12,1): "+str(p.fire([120,235,12,1]))
	print "Result for input vector (40,100,30,1): "+str(p.fire([40,100,30,1]))
	print "Result for input vector (155,55,100,1): "+str(p.fire([155,55,100,1]))
	print "Result for input vector (255,200,55,1): "+str(p.fire([255,200,55,1]))

	data = readFile("rgb.csv")
	# for debugging
	# print data
	t = Trainers(data,0.7)
	#raw_input("Continue?")
	t.gradientDescent(p)
	print "Perceptron's weight vector after training: "+str(p.getWeightVector())

	data = readFile("rgb_test.csv")
	tester(p,data)