单隐层神经网络 python_单隐层神经网络自己代码实现

-- coding: utf-8 --

"""

Created on Fri Sep 21 20:46:51 2018

@author: jianle

"""

import numpy as np

import matplotlib.pyplot as plot

from planar_utils import *

from testCases import *

--------载入数据库

noisy_circles, noisy_moons, blobs, gaussian_quantiles, no_structure=load_extra_datasets()

datasets={"noisy_circles":noisy_circles,

"noisy_moons":noisy_moons,

"blobs": blobs,

"gaussian_quantiles": gaussian_quantiles,

"no_structure": no_structure

}

[X,Y]=datasets["gaussian_quantiles"]

X=X.T

Y.shape=(1,Y.shape[0]) #error 重新定义数组维度

Y=Y.reshape(1,Y.shape[0])

x0=X[0,:]

x1=X[1,:]

-----------绘制数据集-----------------------

plot.scatter(x0,x1,c=np.squeeze(Y),s=60,cmap=plot.cm.Spectral)

-------定义模型每层单元的数量----------------

nx=np.shape(X)[0]

nh=4

ny=np.shape(Y)[0]

----------随机初始化模型参数-----------------

np.random.seed(2)

W1=np.random.randn(nh,nx)0.01

b1=np.zeros((nh,1))

W2=np.random.randn(ny,nh)0.01

b2=np.zeros((ny,1))

parameters={"W1":W1,

"b1":b1,

"W2":W2,

"b2":b2

}

------------test----------------------------

print("W1="+str(W1))

print("b1="+str(b1))

print("W2="+str(W2))

print("b2="+str(b2))

--------------向前传播----------------------

def forward(parameters,X,Y):

W1=parameters["W1"]

b1=parameters["b1"]

W2=parameters["W2"]

b2=parameters["b2"]

Z1=np.dot(W1,X)+b1

A1=np.tanh(Z1)

Z2=np.dot(W2,A1)+b2

A2=sigmoid(Z2)

cache={"Z1":Z1,

"A1":A1,

"Z2":Z2,

"A2":A2

}

return(A2,cache)

--------------向后传播----------------------

def backforward(parameters,cache,X,Y):

W2=parameters["W2"]

A2=cache["A2"]

A1=cache["A1"]

m=np.shape(X)[1]

dZ2=A2-Y

dW2=(1/m)np.dot(dZ2,A1.T)

db2=(1/m)np.sum(dZ2,axis=1,keepdims=True)

dZ1=np.multiply(np.dot(W2.T,dZ2),1-np.power(A1,2))

dW1=(1/m)np.dot(dZ1,X.T)

db1=(1/m)np.sum(dZ1,axis=1,keepdims=True)

grads={"dW1":dW1,

"db1":db1,

"dW2":dW2,

"db2":db2

}

return(grads)

------------更新模型参数---------------

def update(parameters,grads,learning=1.2):

W1=parameters["W1"]

b1=parameters["b1"]

W2=parameters["W2"]

b2=parameters["b2"]

dW1=grads["dW1"]

db1=grads["db1"]

dW2=grads["dW2"]

db2=grads["db2"]

W1=W1-learning*dW1

b1=b1-learning*db1

W2=W2-learning*dW2

b2=b2-learning*db2

#更新

parameters={"W1":W1,

"b1":b1,

"W2":W2,

"b2":b2

}

return (parameters)

运行

Num=8000

for i in range(Num):

A2,cache=forward(parameters,X,Y)

grads=backforward(parameters,cache,X,Y)

parameters=update(parameters,grads,learning=1.2)

def predict(parameters,X):

A2,cache=forward(parameters,X,Y)

predection=np.round(A2)

return predection

-----------绘制预测结果图---------------

plot_decision_boundary(lambda x:predict(parameters,x.T),X,Y)

TIM图片20180921222828.png