libSVM应用举例

前面提到，很多人看到libSVM这么多的参数，估计要犯晕了。没关系，我之前把相关的libSVM参数已经讲解了一遍，这里，再给出libSVM的用法。如果你不想花时间去仔细研究libSVM，完全可以参照我的函数来直接调用libSVM完成你的工作。

首先是训练SVM得到模型；假设，有10个训练样本，每个训练样本，有12个特征值，即：每个训练样本的维数是12，也就是说，训练样本构成了一个10*12的矩阵（当然，所有数据应该被归一化到[-1,1]或[0,1]范围内），另外，所有训练样本，应该有一个明确的类别标签，以此来表明当前样本所属的类别。所以，完整的训练样本，应该是10*13的矩阵，这里，我们给出一个10*13的矩阵，并认为，它就是训练样本。每个训练样本是个行向量，而该矩阵的第一列则代表训练样本的“标签”。即：第一个训练样本，类别标签为“1”，第二个训练样本类别标签为“-1”。第一个训练样本的第一个特征值为0.708333，第二个特征值为1，第三个特征值为1...

doubleinputArr[10][13]=

{

1,0.708333,1,1,-0.320755,-0.105023,-1,1,-0.419847,-1,-0.225806,0,1,

-1,0.583333,-1,0.333333,-0.603774,1,-1,1,0.358779,-1,-0.483871,0,-1,

1,0.166667,1,-0.333333,-0.433962,-0.383562,-1,-1,0.0687023,-1,-0.903226,-1,-1,

-1,0.458333,1,1,-0.358491,-0.374429,-1,-1,-0.480916,1,-0.935484,0,-0.333333,

-1,0.875,-1,-0.333333,-0.509434,-0.347032,-1,1,-0.236641,1,-0.935484,-1,-0.333333,

-1,0.5,1,1,-0.509434,-0.767123,-1,-1,0.0534351,-1,-0.870968,-1,-1,

1,0.125,1,0.333333,-0.320755,-0.406393,1,1,0.0839695,1,-0.806452,0,-0.333333,

1,0.25,1,1,-0.698113,-0.484018,-1,1,0.0839695,1,-0.612903,0,-0.333333,

1,0.291667,1,1,-0.132075,-0.237443,-1,1,0.51145,-1,-0.612903,0,0.333333,

1,0.416667,-1,1,0.0566038,0.283105,-1,1,0.267176,-1,0.290323,0,1

};

另外，我们给出一个待分类的输入向量，并用行向量形式表示：

doubletestArr[]=

{

0.25,1,1,-0.226415,-0.506849,-1,-1,0.374046,-1,-0.83871,0,-1

};

下面，给出完整的SVM训练及预测程序：

#include"stdafx.h"

#include"svm.h"

#include"iostream"

#include"fstream"

usingnamespacestd;

doubleinputArr[10][13]=

{

1,0.708333,1,1,-0.320755,-0.105023,-1,1,-0.419847,-1,-0.225806,0,1,

-1,0.583333,-1,0.333333,-0.603774,1,-1,1,0.358779,-1,-0.483871,0,-1,

1,0.166667,1,-0.333333,-0.433962,-0.383562,-1,-1,0.0687023,-1,-0.903226,-1,-1,

-1,0.458333,1,1,-0.358491,-0.374429,-1,-1,-0.480916,1,-0.935484,0,-0.333333,

-1,0.875,-1,-0.333333,-0.509434,-0.347032,-1,1,-0.236641,1,-0.935484,-1,-0.333333,

-1,0.5,1,1,-0.509434,-0.767123,-1,-1,0.0534351,-1,-0.870968,-1,-1,

1,0.125,1,0.333333,-0.320755,-0.406393,1,1,0.0839695,1,-0.806452,0,-0.333333,

1,0.25,1,1,-0.698113,-0.484018,-1,1,0.0839695,1,-0.612903,0,-0.333333,

1,0.291667,1,1,-0.132075,-0.237443,-1,1,0.51145,-1,-0.612903,0,0.333333,

1,0.416667,-1,1,0.0566038,0.283105,-1,1,0.267176,-1,0.290323,0,1

};

doubletestArr[]=

{

0.25,1,1,-0.226415,-0.506849,-1,-1,0.374046,-1,-0.83871,0,-1

};

voidDefaultSvmParam(structsvm_parameter*param)

{

param->svm_type=C_SVC;

param->kernel_type=RBF;

param->degree=3;

param->gamma=0; //1/num_features

param->coef0=0;

param->nu=0.5;

param->cache_size=100;

param->C=1;

param->eps=1e-3;

param->p=0.1;

param->shrinking=1;

param->probability=0;

param->nr_weight=0;

param->weight_label=NULL;

param->weight=NULL;

}

voidSwitchForSvmParma(structsvm_parameter*param,charch,char*strNum,intnr_fold,intcross_validation)

{

switch(ch)

{

case's':

{

param->svm_type=atoi(strNum);

break;

}

case't':

{

param->kernel_type=atoi(strNum);

break;

}

case'd':

{

param->degree=atoi(strNum);

break;

}

case'g':

{

param->gamma=atof(strNum);

break;

}

case'r':

{

param->coef0=atof(strNum);

break;

}

case'n':

{

param->nu=atof(strNum);

break;

}

case'm':

{

param->cache_size=atof(strNum);

break;

}

case'c':

{

param->C=atof(strNum);

break;

}

case'e':

{

param->eps=atof(strNum);

break;

}

case'p':

{

param->p=atof(strNum);

break;

}

case'h':

{

param->shrinking=atoi(strNum);

break;

}

case'b':

{

param->probability=atoi(strNum);

break;

}

case'q':

{

break;

}

case'v':

{

cross_validation=1;

nr_fold=atoi(strNum);

if(nr_fold<2)

{

cout<<"nr_foldshould>2!!!file:"<<__FILE__<<"function:";

cout<<__FUNCTION__<<"line:"<<__LINE__<<endl;

}

break;

}

case'w':

{

++param->nr_weight;

param->weight_label=(int*)realloc(param->weight_label,sizeof(int)*param->nr_weight);

param->weight=(double*)realloc(param->weight,sizeof(double)*param->nr_weight);

param->weight_label[param->nr_weight-1]=atoi(strNum);

param->weight[param->nr_weight-1]=atof(strNum);

break;

}

default:

{

break;

}

}

}

voidSetSvmParam(structsvm_parameter*param,char*str,intcross_validation,intnr_fold)

{

DefaultSvmParam(param);

cross_validation=0;

charch='';

intstrSize=strlen(str);

for(inti=0;i<strSize;i++)

{

if(str[i]=='-')

{

ch=str[i+1];

intlength=0;

for(intj=i+3;j<strSize;j++)

{

if(isdigit(str[j]))

{

length++;

}

else

{

break;

}

}

char*strNum=newchar[length+1];

intindex=0;

for(intj=i+3;j<i+3+length;j++)

{

strNum[index]=str[j];

index++;

}

strNum[length]='/0';

SwitchForSvmParma(param,ch,strNum,nr_fold,cross_validation);

deletestrNum;

}

}

}

voidSvmTraining(char*option)

{

structsvm_parameterparam;

structsvm_problemprob;

structsvm_model*model;

structsvm_node*x_space;

intcross_validation=0;

intnr_fold=0;

intsampleCount=10;

intfeatureDim=12;

prob.l=sampleCount;

prob.y=newdouble[sampleCount];

prob.x=newstructsvm_node*[sampleCount];

x_space=newstructsvm_node[(featureDim+1)*sampleCount];

SetSvmParam(&param,option,cross_validation,nr_fold);

for(inti=0;i<sampleCount;i++)

{

prob.y[i]=inputArr[i][0];

}

intj=0;

for(inti=0;i<sampleCount;i++)

{

prob.x[i]=&x_space[j];

for(intk=1;k<=featureDim;k++)

{

x_space[i*featureDim+k].index=k;

x_space[i*featureDim+k].value=inputArr[i][k];

}

x_space[(i+1)*featureDim].index=-1;

j=(i+1)*featureDim+1;

}

model=svm_train(&prob,&param);

constchar*model_file_name="C://Model.txt";

svm_save_model(model_file_name,model);

svm_destroy_model(model);

svm_destroy_param(&param);

delete[]prob.y;

delete[]prob.x;

delete[]x_space;

}

intSvmPredict(constchar*modelAdd)

{

structsvm_node*testX;

structsvm_model*testModel;

testModel=svm_load_model(modelAdd);

intfeatureDim=12;

testX=newstructsvm_node[featureDim+1];

for(inti=0;i<featureDim;i++)

{

testX[i].index=i+1;

testX[i].value=testArr[i];

}

testX[featureDim].index=-1;

doublep=svm_predict(testModel,testX);

svm_destroy_model(testModel);

delete[]testX;

if(p>0.5)

{

return1;

}

else

{

return-1;

}

}

int_tmain(intargc,_TCHAR*argv[])

{

SvmTraining("-c100-t1-g4-r1-d4");

intflag=SvmPredict("c://model.txt");

cout<<"flag="<<flag<<endl;

system("pause");

return0;

}