tensorflow前馈神经网络

tensorflow前馈神经网络

# -*- coding: utf-8 -*-
"""
Created on Tue Jul 25 21:58:11 2017

@author: maoyingxue
"""

import tensorflow as tf
import scipy.io as scio
import numpy as np

def load_data(name):
    data=scio.loadmat(name+'.mat')

    data_1 = data[name][2][0]
    data_2 = data[name][2][1]
    data_3 = data[name][2][2]
    data = np.vstack((data_1, data_2, data_3)).astype(np.float32)
    print(data.shape)
    label_1=np.zeros([data_1.shape[0],3])
    label_1[:,0]=1
    #print(label_1)
    label_2=np.zeros([data_2.shape[0],3])
    label_2[:,1]=1
    #print(label_2)
    label_3=np.zeros([data_3.shape[0],3])
    label_3[:,2]=1
    #print(label_3)
    label=np.vstack((label_1,label_2,label_3)).astype(np.float32)
    print(label.shape)
    for i in range(data.shape[0]):
        m=max(data[i])
        data[i]=data[i]/m  
    return data,label

train,train_label=load_data('train') 
test,test_label=load_data('test')     
def add_layer(inputs,in_size,out_size,activation_function=None):
    Weights=tf.Variable(tf.random_normal([in_size,out_size]))
    biases=tf.Variable(tf.zeros([1,out_size])+0.1)
    Wx_plus_b=tf.matmul(inputs,Weights)+biases
    if activation_function is None:
        outputs=Wx_plus_b
    else:
        outputs=activation_function(Wx_plus_b)
    return outputs

def compute_accuracy(v_xs, v_ys):
    global prediction
    y_pre = sess.run(prediction, feed_dict={xs: v_xs})
    correct_prediction = tf.equal(tf.argmax(y_pre,1), tf.argmax(v_ys,1))
    accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
    result = sess.run(accuracy, feed_dict={xs: v_xs, ys: v_ys})
    return result



xs = tf.placeholder(tf.float32, [None, 128]) # 128
ys = tf.placeholder(tf.float32, [None, 3])
l1=add_layer(xs,128,200,activation_function=tf.nn.relu)
prediction=add_layer(l1,200,3,activation_function=tf.nn.softmax)

cross_entropy = tf.reduce_mean(-tf.reduce_sum(ys * tf.log(prediction),
                                              reduction_indices=[1]))       # loss
train_step = tf.train.GradientDescentOptimizer(0.1).minimize(cross_entropy)
sess = tf.Session()
sess.run(tf.global_variables_initializer())
for i in range(6000):
    sess.run(train_step,feed_dict={xs:train,ys:train_label})
    if i%100==0:
        print(compute_accuracy(test,test_label))