python机器学习实现神经网络示例解析

from numpy import array,exp,dot,random

X= array ( [ [0,0,1],[1,1,1],[1,0,1],[0,1,1]])
y = array( [ [0,1,1,0]]).T ## T means "transposition"

x=[-8:0.001:8]
y=1./(1+exp(-x))
plot(x,y)
grid on
text(-6,0.8,['$\frac{1}{1+e^{-x}}$'],'interpreter','latex','fontsize',25)

for it in range(10000):
	z=dot(X,weights)
    output=1/(1+exp(-z))##'dot' play role of "dot product"
    error=y-output
    delta=error*output*(1-output)
    weights+=dot(X.T,delta)

delta=error*output*(1-output)

from numpy import array,exp,dot,random
"""
Created on vscode 10/22/2021
@author Squirre17
"""
X=array([[0,0,1],[1,1,1],[1,0,1],[0,1,1]])
y=array([[0,1,1,0]]).T ## T means "transposition"
random.seed(1)
epochs=10000
weights=2*random.random((3,1))-1## 3 row 1 line, range[-1,1)
for it in range(epochs):
    output=1/(1+exp(-dot(X,weights)))##'dot' play role of "dot product"
    error=y-output
    slope=output*(1-output)
    delta=error*slope
    weights+=dot(X.T,delta)

print(weights)
print(1/(1+exp( -dot([[1,0,0]], weights))))

from numpy import array,exp,dot,random
"""
Created on vscode 10/22/2021
@author Squirre17
"""
X=array([[0,0,1],[0,1,1],[1,0,1],[1,1,1]])
y=array([[0,1,1,0]]).T # T means "transposition"
random.seed(1)
epochs=100000
w0=2*random.random((3,4))-1 # input layer neure
w1=2*random.random((4,1))-1 # hidden layer neure
lr=1
def fp(input):
    l1=1/(1+exp(-dot(input,w0))) # 4×4
    l2=1/(1+exp(-dot(l1,w1))) # 4×1
    return l1,l2
def bp(l1,l2,y):
    l2_error=y-l2
    l2_slope=l2*(1-l2)
    l1_delta=l2_error*l2_slope*lr # 4×1
    l1_error=l1_delta.dot(w1.T)
    l1_slope=l1*(1-l1)
    l0_delta=l1_error*l1_slope*lr
    return l0_delta,l1_delta
for it in range(epochs):
    l0=X
    l1,l2=fp(l0)
    l0_delta,l1_delta=bp(l1,l2,y)
    w1+=dot(l1.T,l1_delta) # 4×4 4×1 # adjust w1 according to loss
    w0+=dot(l0.T,l0_delta)
print(fp([[1,0,0]])[1])