sklearn+MLP多层感知机--训练得到异或门

本博客参考书籍：《scikit-learn机器学习》 [美]加文·海克（Gavin Hackeling）著

神经网络图示

代码实现

from sklearn.model_selection import train_test_split
from sklearn.neural_network import MLPClassifier
y=[0,1,1,0]
x=[[0,0],[0,1],[1,0],[1,1]]

clf=MLPClassifier(solver='lbfgs',activation='logistic',hidden_layer_sizes=(2,),random_state=20)
clf.fit(x,y)

clf.score(x,y)
>>> 1.0

clf.coefs_		# 查看权重矩阵
>>> [array([[6.11804044, 6.35656359],
        [5.79147807, 6.14551866]]),
	 array([[-14.95481488],
	        [ 14.53080958]])]

神经网络正向传播机制验证所得权重

def sigmoid(x):
    return 1/1+np.exp(-x)
w1=np.array([[6.11804044, 6.35656359],
        [5.79147807, 6.14551866]])
w2=np.array([[-14.95481488],
        [ 14.53080958]])
w1.shape
>>> (2,2)
w2.shape
>>> (2,1)
x1=np.array([0,1],ndmin=2)
x1.shape
>>> (1,2)
h_input = np.dot(x1,w1)		# (1,2)(2,2)=(1,2)
h_output = sigmoid(h_input)		# (1,2)

o_input = np.dot(h_output,w2)	# (1,2)(2,1)=(1,1)
o_output=sigmoid(o_input)		# (1,1)
>>> 1.0

看到输出结果我们发现感知机获得的权重参数是正确的！