bpnn神经网络（bp神经网络newff）

本篇文章给大家谈谈bpnn神经网络，以及bp神经网络newff对应的知识点，希望对各位有所帮助，不要忘了收藏本站喔。

本文目录一览：

• 1、BPNN神经网络的Python实现
• 2、如何确定bpnn神经学习的隐含层
• 3、求BP神经网络算法的C++源代码
• 4、BP神经网络流程

BPNN神经网络的Python实现

关于前置的数学知识，可以查看

首先是bpnn.py，渣凯滑实现了BPNeuralNetwork类。

然后是__init__.py，使用mnist手写数字数据集对我们的BPNN进行测试，如腊准确率可达97%以上孙亩。

如何确定bpnn神经学习的隐含层

我自己总结的：

1、神经网络算法隐含层的选取

1.1 构造法

首先运用三种确定隐含层层数的方法得到三个隐含层层数，找到最小值和最大值，然后从最小值开始逐个验证模型预测误差，直到达到最大值。最后选取模型误差最小的那个隐含层层数。该方法适用于双隐含层网络。

1.2 删除法

单隐含层网络非线性映射能力较弱，相族肢同问题，为达到预定映射关系，隐层节点要多一些，以增加网络的可调参数，故适合运用删除法。

1.3黄金分割法

算握键法的主要思想：首先在[a,b]内寻找理想的隐含层节点数，这样就充分保证了网络的逼近能力和泛化能力。为满足高精度逼近的要求，再按照黄金分割原理拓展搜索区间，即得到区间[b,c]（其中b=0.619*（c-a）+a），在区间[b,c]中搜索兆皮世最优，则得到逼近能力更强的隐含层节点数，在实际应用根据要求，从中选取其一即可。

[img]

求BP神经网络算法的C++源代码

// AnnBP.cpp: implementation of the CAnnBP class.

//

///////////////////////////////////////////////////////////////碰枝铅///////

#include "StdAfx.h"

#include "AnnBP.h"

#include "math.h"

/搭扰/////////////////////////////////////////////////////////////////////

// Construction/Destruction

///////////////////////////笑好///////////////////////////////////////////

CAnnBP::CAnnBP()

{

eta1=0.3;

momentum1=0.3;

}

CAnnBP::~CAnnBP()

{

}

double CAnnBP::drnd()

{

return ((double) rand() / (double) BIGRND);

}

/*** 返回-1.0到1.0之间的双精度随机数 ***/

double CAnnBP::dpn1()

{

return (double) (rand())/(32767/2)-1;

}

/*** 作用函数，目前是S型函数 ***/

double CAnnBP::squash(double x)

{

return (1.0 / (1.0 + exp(-x)));

}

/*** 申请1维双精度实数数组 ***/

double* CAnnBP::alloc_1d_dbl(int n)

{

double *new1;

new1 = (double *) malloc ((unsigned) (n * sizeof (double)));

if (new1 == NULL) {

AfxMessageBox("ALLOC_1D_DBL: Couldn't allocate array of doubles\n");

return (NULL);

}

return (new1);

}

/*** 申请2维双精度实数数组 ***/

double** CAnnBP::alloc_2d_dbl(int m, int n)

{

int i;

double **new1;

new1 = (double **) malloc ((unsigned) (m * sizeof (double *)));

if (new1 == NULL) {

AfxMessageBox("ALLOC_2D_DBL: Couldn't allocate array of dbl ptrs\n");

return (NULL);

}

for (i = 0; i m; i++) {

new1[i] = alloc_1d_dbl(n);

}

return (new1);

}

/*** 随机初始化权值 ***/

void CAnnBP::bpnn_randomize_weights(double **w, int m, int n)

{

int i, j;

for (i = 0; i = m; i++) {

for (j = 0; j = n; j++) {

w[i][j] = dpn1();

}

}

}

/*** 0初始化权值 ***/

void CAnnBP::bpnn_zero_weights(double **w, int m, int n)

{

int i, j;

for (i = 0; i = m; i++) {

for (j = 0; j = n; j++) {

w[i][j] = 0.0;

}

}

}

/*** 设置随机数种子 ***/

void CAnnBP::bpnn_initialize(int seed)

{

CString msg,s;

msg="Random number generator seed:";

s.Format("%d",seed);

AfxMessageBox(msg+s);

srand(seed);

}

/*** 创建BP网络 ***/

BPNN* CAnnBP::bpnn_internal_create(int n_in, int n_hidden, int n_out)

{

BPNN *newnet;

newnet = (BPNN *) malloc (sizeof (BPNN));

if (newnet == NULL) {

printf("BPNN_CREATE: Couldn't allocate neural network\n");

return (NULL);

}

newnet-input_n = n_in;

newnet-hidden_n = n_hidden;

newnet-output_n = n_out;

newnet-input_units = alloc_1d_dbl(n_in + 1);

newnet-hidden_units = alloc_1d_dbl(n_hidden + 1);

newnet-output_units = alloc_1d_dbl(n_out + 1);

newnet-hidden_delta = alloc_1d_dbl(n_hidden + 1);

newnet-output_delta = alloc_1d_dbl(n_out + 1);

newnet-target = alloc_1d_dbl(n_out + 1);

newnet-input_weights = alloc_2d_dbl(n_in + 1, n_hidden + 1);

newnet-hidden_weights = alloc_2d_dbl(n_hidden + 1, n_out + 1);

newnet-input_prev_weights = alloc_2d_dbl(n_in + 1, n_hidden + 1);

newnet-hidden_prev_weights = alloc_2d_dbl(n_hidden + 1, n_out + 1);

return (newnet);

}

/* 释放BP网络所占地内存空间 */

void CAnnBP::bpnn_free(BPNN *net)

{

int n1, n2, i;

n1 = net-input_n;

n2 = net-hidden_n;

free((char *) net-input_units);

free((char *) net-hidden_units);

free((char *) net-output_units);

free((char *) net-hidden_delta);

free((char *) net-output_delta);

free((char *) net-target);

for (i = 0; i = n1; i++) {

free((char *) net-input_weights[i]);

free((char *) net-input_prev_weights[i]);

}

free((char *) net-input_weights);

free((char *) net-input_prev_weights);

for (i = 0; i = n2; i++) {

free((char *) net-hidden_weights[i]);

free((char *) net-hidden_prev_weights[i]);

}

free((char *) net-hidden_weights);

free((char *) net-hidden_prev_weights);

free((char *) net);

}

/*** 创建一个BP网络，并初始化权值***/

BPNN* CAnnBP::bpnn_create(int n_in, int n_hidden, int n_out)

{

BPNN *newnet;

newnet = bpnn_internal_create(n_in, n_hidden, n_out);

#ifdef INITZERO

bpnn_zero_weights(newnet-input_weights, n_in, n_hidden);

#else

bpnn_randomize_weights(newnet-input_weights, n_in, n_hidden);

#endif

bpnn_randomize_weights(newnet-hidden_weights, n_hidden, n_out);

bpnn_zero_weights(newnet-input_prev_weights, n_in, n_hidden);

bpnn_zero_weights(newnet-hidden_prev_weights, n_hidden, n_out);

return (newnet);

}

void CAnnBP::bpnn_layerforward(double *l1, double *l2, double **conn, int n1, int n2)

{

double sum;

int j, k;

/*** 设置阈值 ***/

l1[0] = 1.0;

/*** 对于第二层的每个神经元 ***/

for (j = 1; j = n2; j++) {

/*** 计算输入的加权总和 ***/

sum = 0.0;

for (k = 0; k = n1; k++) {

sum += conn[k][j] * l1[k];

}

l2[j] = squash(sum);

}

}

/* 输出误差 */

void CAnnBP::bpnn_output_error(double *delta, double *target, double *output, int nj, double *err)

{

int j;

double o, t, errsum;

errsum = 0.0;

for (j = 1; j = nj; j++) {

o = output[j];

t = target[j];

delta[j] = o * (1.0 - o) * (t - o);

errsum += ABS(delta[j]);

}

*err = errsum;

}

/* 隐含层误差 */

void CAnnBP::bpnn_hidden_error(double *delta_h, int nh, double *delta_o, int no, double **who, double *hidden, double *err)

{

int j, k;

double h, sum, errsum;

errsum = 0.0;

for (j = 1; j = nh; j++) {

h = hidden[j];

sum = 0.0;

for (k = 1; k = no; k++) {

sum += delta_o[k] * who[j][k];

}

delta_h[j] = h * (1.0 - h) * sum;

errsum += ABS(delta_h[j]);

}

*err = errsum;

}

/* 调整权值 */

void CAnnBP::bpnn_adjust_weights(double *delta, int ndelta, double *ly, int nly, double **w, double **oldw, double eta, double momentum)

{

double new_dw;

int k, j;

ly[0] = 1.0;

for (j = 1; j = ndelta; j++) {

for (k = 0; k = nly; k++) {

new_dw = ((eta * delta[j] * ly[k]) + (momentum * oldw[k][j]));

w[k][j] += new_dw;

oldw[k][j] = new_dw;

}

}

}

/* 进行前向运算 */

void CAnnBP::bpnn_feedforward(BPNN *net)

{

int in, hid, out;

in = net-input_n;

hid = net-hidden_n;

out = net-output_n;

/*** Feed forward input activations. ***/

bpnn_layerforward(net-input_units, net-hidden_units,

net-input_weights, in, hid);

bpnn_layerforward(net-hidden_units, net-output_units,

net-hidden_weights, hid, out);

}

/* 训练BP网络 */

void CAnnBP::bpnn_train(BPNN *net, double eta, double momentum, double *eo, double *eh)

{

int in, hid, out;

double out_err, hid_err;

in = net-input_n;

hid = net-hidden_n;

out = net-output_n;

/*** 前向输入激活 ***/

bpnn_layerforward(net-input_units, net-hidden_units,

net-input_weights, in, hid);

bpnn_layerforward(net-hidden_units, net-output_units,

net-hidden_weights, hid, out);

/*** 计算隐含层和输出层误差 ***/

bpnn_output_error(net-output_delta, net-target, net-output_units,

out, out_err);

bpnn_hidden_error(net-hidden_delta, hid, net-output_delta, out,

net-hidden_weights, net-hidden_units, hid_err);

*eo = out_err;

*eh = hid_err;

/*** 调整输入层和隐含层权值 ***/

bpnn_adjust_weights(net-output_delta, out, net-hidden_units, hid,

net-hidden_weights, net-hidden_prev_weights, eta, momentum);

bpnn_adjust_weights(net-hidden_delta, hid, net-input_units, in,

net-input_weights, net-input_prev_weights, eta, momentum);

}

/* 保存BP网络 */

void CAnnBP::bpnn_save(BPNN *net, char *filename)

{

CFile file;

char *mem;

int n1, n2, n3, i, j, memcnt;

double dvalue, **w;

n1 = net-input_n; n2 = net-hidden_n; n3 = net-output_n;

printf("Saving %dx%dx%d network to '%s'\n", n1, n2, n3, filename);

try

{

file.Open(filename,CFile::modeWrite|CFile::modeCreate|CFile::modeNoTruncate);

}

catch(CFileException* e)

{

e-ReportError();

e-Delete();

}

file.Write(n1,sizeof(int));

file.Write(n2,sizeof(int));

file.Write(n3,sizeof(int));

memcnt = 0;

w = net-input_weights;

mem = (char *) malloc ((unsigned) ((n1+1) * (n2+1) * sizeof(double)));

// mem = (char *) malloc (((n1+1) * (n2+1) * sizeof(double)));

for (i = 0; i = n1; i++) {

for (j = 0; j = n2; j++) {

dvalue = w[i][j];

//fastcopy(mem[memcnt], dvalue, sizeof(double));

fastcopy(mem[memcnt], dvalue, sizeof(double));

memcnt += sizeof(double);

}

}

file.Write(mem,sizeof(double)*(n1+1)*(n2+1));

free(mem);

memcnt = 0;

w = net-hidden_weights;

mem = (char *) malloc ((unsigned) ((n2+1) * (n3+1) * sizeof(double)));

// mem = (char *) malloc (((n2+1) * (n3+1) * sizeof(double)));

for (i = 0; i = n2; i++) {

for (j = 0; j = n3; j++) {

dvalue = w[i][j];

fastcopy(mem[memcnt], dvalue, sizeof(double));

// fastcopy(mem[memcnt], dvalue, sizeof(double));

memcnt += sizeof(double);

}

}

file.Write(mem, (n2+1) * (n3+1) * sizeof(double));

// free(mem);

file.Close();

return;

}

/* 从文件中读取BP网络 */

BPNN* CAnnBP::bpnn_read(char *filename)

{

char *mem;

BPNN *new1;

int n1, n2, n3, i, j, memcnt;

CFile file;

try

{

file.Open(filename,CFile::modeRead|CFile::modeCreate|CFile::modeNoTruncate);

}

catch(CFileException* e)

{

e-ReportError();

e-Delete();

}

// printf("Reading '%s'\n", filename);// fflush(stdout);

file.Read(n1, sizeof(int));

file.Read(n2, sizeof(int));

file.Read(n3, sizeof(int));

new1 = bpnn_internal_create(n1, n2, n3);

// printf("'%s' contains a %dx%dx%d network\n", filename, n1, n2, n3);

// printf("Reading input weights..."); // fflush(stdout);

memcnt = 0;

mem = (char *) malloc (((n1+1) * (n2+1) * sizeof(double)));

file.Read(mem, ((n1+1)*(n2+1))*sizeof(double));

for (i = 0; i = n1; i++) {

for (j = 0; j = n2; j++) {

//fastcopy((new1-input_weights[i][j]), mem[memcnt], sizeof(double));

fastcopy((new1-input_weights[i][j]), mem[memcnt], sizeof(double));

memcnt += sizeof(double);

}

}

free(mem);

// printf("Done\nReading hidden weights..."); //fflush(stdout);

memcnt = 0;

mem = (char *) malloc (((n2+1) * (n3+1) * sizeof(double)));

file.Read(mem, (n2+1) * (n3+1) * sizeof(double));

for (i = 0; i = n2; i++) {

for (j = 0; j = n3; j++) {

//fastcopy((new1-hidden_weights[i][j]), mem[memcnt], sizeof(double));

fastcopy((new1-hidden_weights[i][j]), mem[memcnt], sizeof(double));

memcnt += sizeof(double);

}

}

free(mem);

file.Close();

printf("Done\n"); //fflush(stdout);

bpnn_zero_weights(new1-input_prev_weights, n1, n2);

bpnn_zero_weights(new1-hidden_prev_weights, n2, n3);

return (new1);

}

void CAnnBP::CreateBP(int n_in, int n_hidden, int n_out)

{

net=bpnn_create(n_in,n_hidden,n_out);

}

void CAnnBP::FreeBP()

{

bpnn_free(net);

}

void CAnnBP::Train(double *input_unit,int input_num, double *target,int target_num, double *eo, double *eh)

{

for(int i=1;i=input_num;i++)

{

net-input_units[i]=input_unit[i-1];

}

for(int j=1;j=target_num;j++)

{

net-target[j]=target[j-1];

}

bpnn_train(net,eta1,momentum1,eo,eh);

}

void CAnnBP::Identify(double *input_unit,int input_num,double *target,int target_num)

{

for(int i=1;i=input_num;i++)

{

net-input_units[i]=input_unit[i-1];

}

bpnn_feedforward(net);

for(int j=1;j=target_num;j++)

{

target[j-1]=net-output_units[j];

}

}

void CAnnBP::Save(char *filename)

{

bpnn_save(net,filename);

}

void CAnnBP::Read(char *filename)

{

net=bpnn_read(filename);

}

void CAnnBP::SetBParm(double eta, double momentum)

{

eta1=eta;

momentum1=momentum;

}

void CAnnBP::Initialize(int seed)

{

bpnn_initialize(seed);

}

BP神经网络流程

表示m+1层网络，第i层节点数桐扰知为 ，输入为 ，输入为

weight[m]=

threshold[m]=

delta_weight[m]=

delta_threshold[m]=

输局消入为 ，输入为

for k in range(1,m+1):#循环m次

----#

----#

初始化：

E_H[m]=

H_I[m]=

for k in range(m,0,-1):

----#1.

----#2.

----#3.

----#4.

weight[m]=weight[m]+delta_weight[m]

threshold[m]=threshold[m]+delta_threshold[m]

python实现的BP神经网李胡络

关于bpnn神经网络和bp神经网络newff的介绍到此就结束了，不知道你从中找到你需要的信息了吗 ？如果你还想了解更多这方面的信息，记得收藏关注本站。