《统计学习方法》-感知机-例2.1

刚开始学习机器学习，试着将例子用程序写出来，写的有点low，欢迎大家吐槽！

import numpy as np

x = np.array([[3,3],[4,3],[1,1]])    #[[3,3],[4,3],[1,1]]
y = np.array([1,1,-1])

w0 = np.array([0,0])
b0 = 0

lr = 1
loss = 0

print(w0,b0)
i = 0
count = 0 
while True:
    count += 1
    if i > 2:
        i = i % 2 - 1
        print(i)
        
    if y[i] * (w0.dot(x[i].T) + b0) <= 0:
        print('error '+str(i))
        w0 = w0 + y[i] * x[i]
        b0 = b0 + y[i]
        i = 0
        print('w0 = '+str(w0),'b0 = '+str(b0))
    else:
        print('coorct '+str(i))
        i += 1
        if i == 3:
            print('w0 = '+str(w0),'b0 = '+str(b0))
    
    if count > 30:
        break

[0 0] 0
error 0
w0 = [3 3] b0 = 1
coorct 0
coorct 1
error 2
w0 = [2 2] b0 = 0
coorct 0
coorct 1
error 2
w0 = [1 1] b0 = -1
coorct 0
coorct 1
error 2
w0 = [0 0] b0 = -2
error 0
w0 = [3 3] b0 = -1
coorct 0
coorct 1
error 2
w0 = [2 2] b0 = -2
coorct 0
coorct 1
error 2
w0 = [1 1] b0 = -3
coorct 0
coorct 1
coorct 2
w0 = [1 1] b0 = -3
0
coorct 0
coorct 1
coorct 2
w0 = [1 1] b0 = -3
0
coorct 0
coorct 1
coorct 2
w0 = [1 1] b0 = -3
0
coorct 0
coorct 1
coorct 2
w0 = [1 1] b0 = -3
0
coorct 0
coorct 1

感知机学习算法由于采用不同的初值或选取不同的误分类点，解可以不同！