转载 libsvm vc 移植 实现多类分类

http://zllxsha.blog.163.com/blog/static/50555091201010981932412/ by piglet

在网上找了很久关于libsvm 在vc上的移植，网上大部分讲的都是两类的分类

在此分享下，我实现了多类的分类

代码对于提取的数据，可能在部分的地方需要修改，比如样本的个数等

当然，工程中需要包含svm.h,svm.cpp

代码如下：

#include <stdio.h>
#include <ctype.h>
#include <list>
#include "svm.h"    //包涵本目录下的SVM.H头文件
#include <fstream>
#include <iostream>
#include <string>
    
using namespace std;
void main()
{
    /*
    struct svm_problem
    {
        int l;
        double *y;
        struct svm_node **x;
    };
    struct svm_node
    {
        int index;
        double value;
    };
    */
    ifstream inFile;
    ifstream inFile1;
 ofstream outFile("out.txt");
 ofstream outFile1("out2.txt");
    svm_problem prob;                    //建立SVM数据
    prob.l=2262;                            //80个样本，前40个为正样本，train.1.txt=3089,train.2.txt = 2262
    double d[2262];
    int probfeature=6;                    //样本特征维数
    prob.y=new double[prob.l];
    svm_parameter param;
    
    //svm参数设置
    param.svm_type = C_SVC;
    param.kernel_type = RBF;
    param.degree = 3;
    param.gamma = 0.0001;
    param.coef0 = 0;
    param.nu = 0.5;
    param.cache_size = 100;
    param.C = 10;
    param.eps = 1e-5;
    param.p = 0.1;
    param.shrinking = 1;
    param.probability = 0;
    param.nr_weight = 0;
    param.weight_label = NULL;
    param.weight = NULL;

    
        //分类
        if(param.gamma == 0) param.gamma = 0.5;                //gamma for  poly/rbf/sigmoid
        svm_node *x_space = new svm_node[(probfeature+1) * prob.l];       //样本特征存储空间
        prob.x = new svm_node *[prob.l];                    //每一个X指向一个样本
        //数据填充开始
         inFile.open("train.2.txt");                            //打开训练数据文件
        
        for(int i=0;i<prob.l;i++)//prob.l 总样本数
        {
        /**/
            float fvalue;
   int cn,cn2;
   char ch;
   inFile>>cn;
   outFile<<cn<<" ";
            for (int j=0;j<probfeature;j++)
            {
    
                inFile>>cn2>>ch>>fvalue;
                if(fvalue!=0.0)
                {
                    x_space[(probfeature+1)*i+j].index=j+1;
                    x_space[(probfeature+1)*i+j].value=fvalue;
                }
    outFile<<cn2<<ch<<fvalue<<" ";
            }
            x_space[(probfeature+1)*i+probfeature].index=-1;
            prob.x[i]=&x_space[(probfeature+1)*i];
   prob.y[i] = cn;
   outFile<<prob.y[i]<<endl;
   
   //         if (i<2000)                                        //类别标签，两类的时候
   //{
   //             prob.y[i]=1;
   // outFile<<"1"<<endl;
   //}
   //         else
   //{
   //             prob.y[i]=-1;
   // outFile<<"-1"<<endl;
   //}
        }
  outFile.close();
        inFile.close();                                        //关闭文件

        //数据填充结束
        // build model & classify
        svm_model *model = svm_train(&prob, &param);    //开始训练
        int test_num=1109;                              //test.1.txt = 4000; test.2.txt = 1109
        inFile1.open("test.2.txt");
        svm_node x[7];                                    //定义一个测试样本
  double d2[1109];
  int right=0;
    
        for(int i1=0;i1<test_num;i1++)
        { 
            float fvalue;
   int cn,cn2;
   char ch;
   inFile1>>cn;
   outFile1<<cn<<" ";
            for (int j=0;j<probfeature;j++)
            {
                inFile1>>cn2>>ch>>fvalue;
    //cout<<"fvalue: "<<fixed<<fvalue<<endl;
                if(fvalue!=0.0)
                {
                    x[j].index=j+1;
                    x[j].value=fvalue;
     //cout<<x[j].index<<"   "<<x[j].value<<endl;
                }
    outFile1<<cn2<<ch<<fvalue<<" ";
            }    
            x[7].index=-1;////////////********************为什么要多一个值，而且还是空的*****************？？？？？//////
   
         d2[i1]=svm_predict(model,x);
   //cout<<"cn: "<<cn<<"  d2[i1]: "<<d2[i1]<<"  "<<endl;
   outFile1<<d2[i1]<<endl;
   if(d2[i1] == cn)
    right++;
   //if(cn==1&&d2[i1]==1)//当2类的时候，判断分类
   // right++;
   //if(cn==0&&d2[i1]==-1)
   // right++;
        }
        inFile1.close();
  //outFile1.close();
  x[0].index=1;
  x[0].value = 0.2;
  x[1].index=2;
  x[1].value = 0.2;
  x[2].index=3;
  x[2].value = 0.2;
  x[3].index=4;
  x[3].value = 0.2;

  cout<<"svm_predict(model,x): "<<svm_predict(model,x)<<endl;
        svm_destroy_model(model);
        delete[] x_space;
        delete[] prob.x;
        delete[] prob.y;
  cout<<"accuracy: "<<right<<"/"<<test_num<<"    "<<(float)right/(float)test_num<<endl;

}