神经网络C语言实现(共4页).docx

《神经网络C语言实现(共4页).docx》由会员分享，可在线阅读，更多相关《神经网络C语言实现(共4页).docx（4页珍藏版）》请在淘文阁 - 分享文档赚钱的网站上搜索。

1、精选优质文档-倾情为你奉上#include stdio.h#include const double e = 2.;/设置一个神经网络/有一个隐藏层（含有两个节点）/输出层有一个节点/输入数据是二维（两个节点）/一个样本数据为：x = (0.35,0.9) 标签为0.5/初始权值输入节点1到隐藏层：0.1，0.4/输入节点2到隐藏层：0.8,0.6/隐藏层到输出层初始权值为：0.3,0.9/学习速率为1double changeWeightFromHiddenToOutput(double cost,double output,double hiddenLayerCode)double res

2、ult=0;result = cost*output*(1-output)*hiddenLayerCode;return result;double changeWeightFromInputToHidden(double cost,double output,double weightOfHiddenCodeToOutput,double weightOfHiddenCode,double inputNum)double result=0;result = cost*output*(1-output)*weightOfHiddenCodeToOutput*weightOfHiddenCode

3、*(1-weightOfHiddenCode)*inputNum;return result;double sigmoidFunction(double x)double result=0;result = 1/(1+pow(e,-x);return result;double costFunction(double originalSignal,double outputOfOurCalculation)/此处采取的损失函数是最小二乘法double cost=0;cost = (1/2.0)*(originalSignal-outputOfOurCalculation)*(originalS

4、ignal-outputOfOurCalculation);return cost;double upDateWeightFunction(double originalValue,double gradient)double updatedWeight=originalValue;updatedWeight = updatedWeight - fabs(gradient);return updatedWeight;int main(void)double weightFromInputToHidden2=0.1,0.4,0.8,0.6;double weightFromHiddenToOut

5、put=0.3,0.9;double firstHiddenCode,secondHiddenCode,outputCode;double inputValue =0.35,0.9;double originalSignal = 0.5;double cost=0;double weightChangeNum=0;double addWeightSum = 0;firstHiddenCode = 0;secondHiddenCode = 0;outputCode = 0;/前向传播addWeightSum = weightFromInputToHidden00*inputValue0 + we

6、ightFromInputToHidden10*inputValue1; firstHiddenCode = sigmoidFunction(addWeightSum);addWeightSum = weightFromInputToHidden01*inputValue0 + weightFromInputToHidden11*inputValue1;secondHiddenCode = sigmoidFunction(addWeightSum);addWeightSum = weightFromHiddenToOutput0*firstHiddenCode + weightFromHidd

7、enToOutput1*secondHiddenCode; outputCode = sigmoidFunction(addWeightSum);/计算误差cost = costFunction(originalSignal,outputCode);printf(nn(0)firNode:%f secNode:%f outNode:%f cost:%f,firstHiddenCode,secondHiddenCode,outputCode,cost);printf(nntt输入到隐藏层的权值:tt);for(int i=0;i2;i+)printf(ntt);for(int j=0;j2;j+

8、)printf( %lf ,weightFromInputToHiddenij);printf(nntt隐藏层到输出的权值:ntt);for(i=0;i2;i+)printf( %lf ,weightFromHiddenToOutputi);for(int iteration = 0;iteration1;iteration+)/更新隐藏层到输出层权值/weightChangeNum为相应权值的梯度weightChangeNum = changeWeightFromHiddenToOutput(cost,outputCode,firstHiddenCode);weightFromHiddenT

9、oOutput0 = upDateWeightFunction(weightFromHiddenToOutput0,weightChangeNum);weightChangeNum = changeWeightFromHiddenToOutput(cost,outputCode,secondHiddenCode);weightFromHiddenToOutput1 = upDateWeightFunction(weightFromHiddenToOutput1,weightChangeNum);/更新输入层到隐藏层的权值weightChangeNum = changeWeightFromInp

10、utToHidden(cost,outputCode,weightFromHiddenToOutput0,firstHiddenCode,inputValue0);weightFromInputToHidden00 = upDateWeightFunction(weightFromInputToHidden00,weightChangeNum);weightChangeNum = changeWeightFromInputToHidden(cost,outputCode,weightFromHiddenToOutput1,secondHiddenCode,inputValue0);weight

11、FromInputToHidden01 = upDateWeightFunction(weightFromInputToHidden01,weightChangeNum);weightChangeNum = changeWeightFromInputToHidden(cost,outputCode,weightFromHiddenToOutput0,firstHiddenCode,inputValue1);weightFromInputToHidden10 = upDateWeightFunction(weightFromInputToHidden10,weightChangeNum);wei

12、ghtChangeNum = changeWeightFromInputToHidden(cost,outputCode,weightFromHiddenToOutput1,secondHiddenCode,inputValue1);weightFromInputToHidden11 = upDateWeightFunction(weightFromInputToHidden11,weightChangeNum);/再次进行前向传播addWeightSum = weightFromInputToHidden00*inputValue0 + weightFromInputToHidden10*i

13、nputValue1; firstHiddenCode = sigmoidFunction(addWeightSum);addWeightSum = weightFromInputToHidden01*inputValue0 + weightFromInputToHidden11*inputValue1;secondHiddenCode = sigmoidFunction(addWeightSum);/输出addWeightSum = weightFromHiddenToOutput0*firstHiddenCode + weightFromHiddenToOutput1*secondHidd

14、enCode; outputCode = sigmoidFunction(addWeightSum);/计算误差cost = costFunction(originalSignal,outputCode);printf(nn(%d)firNode:%f secNode:%f outNode:%f cost:%f,iteration+1,firstHiddenCode,secondHiddenCode,outputCode,cost);printf(nntt输入到隐藏层的权值:tt);for(int i=0;i2;i+)printf(ntt);for(int j=0;j2;j+)printf( %lf ,weightFromInputToHiddenij);printf(nntt隐藏层到输出的权值:ntt);for(i=0;i2;i+)printf( %lf ,weightFromHiddenToOutputi);if(cost0.01)printf(nn误差已经小于0.01了！停止_n);break;printf(nntt); 专心-专注-专业