NeuralNetworksInPython/SingleSigmoidNeuron.py at master · zbkruturaj/NeuralNetworksInPython · 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
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
import numpy as np
import pandas as pd
import random

def eval_perceptron(x,w):
	return 1 if sigmoid(np.dot(x,w))>=0.5 else 0

def sigmoid(x):
	return 1 / float(1 + np.exp(- x))

def conv(df, head):
	df[head] = df[head].astype('category')
	df[head].cat.categories = range(len(df[head].cat.categories))
	df[head] = df[head].astype('float64')

##
headers = ["age", "job", "marital", "education", "default", "balance", "housing", "loan", "contact", "day", "month", "duration", "campaign", "pdays", "previous", "poutcome", "y"]
df = pd.read_csv('bank.csv', names=headers)
total_records = len(df)
c_h = ["job", "marital", "education", "default", "housing", "loan", "contact", "month", "duration", "previous", "poutcome"]
map(lambda x:conv(df,x), c_h)
df["y"] = df["y"].astype('category')
df["y"].cat.categories = [0,1]
df["y"] = df["y"].astype('float64')
msk = np.random.randn(total_records) <= 0.5

df_tr = df.sample(frac=0.5,random_state=627)
df_te = df.drop(df_tr.index)

tar_tr = df_tr['y']
tar_te = df_te['y']


del df_tr['y']
del df_te['y']


## l = df_tr.values.tolist()
## tl = tar_tr.tolist()

## print l[0]
## print tl[0]
d = (df_tr - df_tr.mean() / (df_tr.max()-df_tr.min()))
l_tr = d.values.tolist()
t_tr = tar_tr.tolist()

d = (df_te - df_te.mean() / (df_te.max()-df_te.min()))
l_te = d.values.tolist()
t_te = tar_te.tolist()
## print d.head()

inputs = l_tr
outputs = t_tr
learning_rate = 0.3
n_inputs = len(inputs[0])
epochs = 1000
w = []
r = random.Random(1221)#2612)#2111)
for i in range(n_inputs):
    w.append(r.random()*2-1)
print w
w = np.asarray(w)
score = 0.0
for (inp,out) in zip(l_te,t_te):
	if out == eval_perceptron(inp,w):
		score += 1.0

print score*100/len(t_te)

for i in range(epochs):
    for (inp,out) in zip(inputs,outputs):
	inp = np.asarray(inp)
        w = w - (sigmoid(np.dot(w,inp))-out)*(sigmoid(np.dot(w,inp))*(1-sigmoid(np.dot(w,inp))))*inp
print w

score = 0.0
for (inp,out) in zip(l_te,t_te):
	if out == eval_perceptron(inp,w):
		score += 1.0

print score*100/len(t_te)

score = 0.0
for (inp,out) in zip(l_tr,t_tr):
	if out == eval_perceptron(inp,w):
		score += 1.0

print score*100/len(t_tr)