import numpy as np
class Logistic():
def __init__(self):
pass
def sigmoid(self, z):
'''激活函数'''
return 1 / (1 + np.exp(-z))
def logistic(self, X, theta):
'''一层神经网络进行简单的逻辑运算'''
h = self.sigmoid(X*theta.T)
for i in range(int(h.shape[1])):
h[0, i] = 1 if h[0, i] >= 0.5 else 0
return h
X = np.matrix(np.array([1, 0, 1]))
theta1 = np.matrix([[-30, 20, 20], [10, -20, -20]]) #第一层网络的权重
theta2 = np.matrix([-10, 20, -20]) #第二层网络的权重
log = Logistic() #实例化
a1 = log.logistic(X, theta1) #第一层
a1 = np.c_[1, a1] #添加偏置单元
a2 = log.logistic(a1, theta2) #第二层
print(a2)
用神经网络实现异或运算
最新推荐文章于 2024-08-14 11:48:20 发布