深度学习之感知器

import numpy as np

x1 = np.array([0, 0, 1, 1])
x2 = np.array([0, 1, 0, 1])
y1 = np.array([0, 0, 0, 1])
y2 = np.array([0, 1, 1, 1]) #真值表

alpha = 0.01
w0 = -0.8
w1 = 0.6
w2 = 0.4

def fx(x1, x2):
    return w1*x1+w2*x2+w0    #定义阶跃函数

def Fx(m,n):        #激活函数
    if(fx(m,n)>0):
        return 1
    else:
        return 0

#与运算循环
for i in range(5000):
    w0 = w0 - (fx(x1[i % 4], x2[i % 4]) - y1[i % 4]) * alpha
    w1 = w1 - (fx(x1[i % 4], x2[i % 4]) - y1[i % 4]) * x1[i % 4] * alpha
    w2 = w2 - (fx(x1[i % 4], x2[i % 4]) - y1[i % 4]) * x2[i % 4] * alpha
    if (abs(fx(x1[i % 4], x2[i % 4]) - y1[i % 4]) < 0.001 and abs(      # 如果偏导小于0.001,输出此时权重及偏导
            (fx(x1[i % 4], x2[i % 4]) - y1[i % 4]) * x1[i % 4]) < 0.001 and abs(
        (fx(x1[i % 4], x2[i % 4]) - y1[i % 4]) * x2[i % 4]) < 0.01):
        print(w0, w1, w2)
        break

# 或运算
for i in range(5000):
    w0 = w0 - (fx(x1[i % 4], x2[i % 4]) - y2[i % 4]) * alpha
    w1 = w1 - (fx(x1[i % 4], x2[i % 4]) - y2[i % 4]) * x1[i % 4] * alpha
    w2 = w2 - (fx(x1[i % 4], x2[i % 4]) - y2[i % 4]) * x2[i % 4] * alpha
    if (abs(fx(x1[i % 4], x2[i % 4]) - y2[i % 4]) < 0.001 and abs(
            (fx(x1[i % 4], x2[i % 4]) - y2[i % 4]) * x1[i % 4]) < 0.001 and abs(
        (fx(x1[i % 4], x2[i % 4]) - y2[i % 4]) * x2[i % 4]) < 0.01):
        print(w0, w1, w2)
        break