classificationldf.cpp

该程序包实现了几个常用的模式识别分类器算法

/* CLdf类实现线性判别方程算法（LDF）。
   使用方法如下：创建一个对象，调用train方法训练分类器，调用test方法进行分类。 */
#include "stdlib.h"
#include <iostream>
#include <string.h>
#include <fstream>
#include <math.h>
using namespace std;
#include "global.h"

CLdf::CLdf()
{
	paramU = NULL;
	paramSigma0 = NULL;
	paramW = NULL;
	paramW0 = NULL;
	resultTotal = NULL;
	resultRight = NULL;
}

CLdf::~CLdf()
{
	if (this->paramU != NULL)
		delete[] this->paramU;
	if (this->paramSigma0 != NULL)
		delete[] this->paramSigma0;
	if (this->paramW != NULL)
		delete[] this->paramW;
	if (this->paramW0 != NULL)
		delete[] this->paramW0;
	if (this->resultTotal != NULL)
		delete[] this->resultTotal;
	if (this->resultRight != NULL)
		delete[] this->resultRight;
}

/* 训练分类器，输入训练样本的文件名，返回true表示训练成功，返回false表示训练失败。 */
bool CLdf::train(char* fileName)
{
	bool ret = true;
	int i, j, k, index, cnum;
	ret = this->sdata.readFile(fileName);//读入训练数据
	this->sdata.normalize();
	if (!ret)
		return false;//如果数据格式不正确，退出程序。
	int numClass = this->sdata.numClass;
	int numFeature = this->sdata.numFeature;
	int numSample = this->sdata.numSample;
	DOUBLE *sumX = new DOUBLE[numClass*numFeature];
	DOUBLE *sumXX = new DOUBLE[numFeature*numFeature];
	int* classSample = new int[numClass];
	int total;
	for (i=0; i<numClass; i++)
	{
		classSample[i]=0;
	}
	total = numClass * numFeature;
	for (i=0; i<total; i++)
	{
		sumX[i]=0;
	}
	total = numFeature * numFeature;
	for (i=0; i<total; i++)
	{
		sumXX[i]=0;
	}

	for (i=0; i<numSample; i++)
	{
		cnum = this->sdata.ydata[i];
		classSample[cnum] ++;
		for (j=0; j<numFeature; j++)
		{
			sumX[cnum*numFeature+j] += this->sdata.xdata[i*numFeature+j];
		}
	}
	for (j=0; j<numFeature; j++)
	{
		for (k=j; k<numFeature; k++)
		{
			for (i=0; i<numSample; i++)
			{
				sumXX[j*numFeature+k] +=this->sdata.xdata[i*numFeature+j] * this->sdata.xdata[i*numFeature+k];
			}
		}
	}
	this->paramU = new DOUBLE[numClass*numFeature];
	this->paramSigma0 = new DOUBLE[numFeature*numFeature];
	int totalSample = this->sdata.numSample;
	for (i=0; i<numClass; i++)
	{
		for (j=0; j<numFeature; j++)
		{
			this->paramU[i*numFeature+j] = sumX[i*numFeature+j] / classSample[i];
		}
	}
	for (j=0; j<numFeature; j++)
	{
		for (k=j; k<numFeature; k++)
		{
			DOUBLE sigma = sumXX[j*numFeature+k];
			for (i=0; i<numClass; i++)
			{
				sigma -= this->paramU[i*numFeature+j]*this->paramU[i*numFeature+k]*classSample[i];
			}
			this->paramSigma0[j*numFeature+k] = sigma/totalSample;
		}
	}
	for (j=1; j<numFeature; j++)
	{
		for (k=0; k<j; k++)
		{
			this->paramSigma0[j*numFeature+k] = this->paramSigma0[k*numFeature+j];
		}
	}
	for (j=0; j<numFeature; j++)//避免方差为零的问题
	{
		//cout<<"对角："<<this->paramSigma0[j*numFeature+j];
		if (this->paramSigma0[j*numFeature+j]<0.001)
		{
			this->paramSigma0[j*numFeature+j] += 0.001;
		}
		//cout<<" "<<this->paramSigma0[j*numFeature+j]<<endl;
	}
	//计算W和W0参数
	this->paramW = new DOUBLE[numClass * numFeature];
	this->paramW0 = new DOUBLE[numClass];
	CMatrix sigma0;
	CMatrix u;
	CMatrix invSigma0;
	CMatrix transU;
	sigma0.nline = numFeature;
	sigma0.ncol = numFeature;
	sigma0.pdata = new DOUBLE[sigma0.nline * sigma0.ncol];
	for (i=0; i<(numFeature*numFeature); i++)
	{
		sigma0.pdata[i] = this->paramSigma0[i];
	}
	//sigma0.shownumber();
	sigma0.inv(invSigma0);//矩阵求逆
	u.nline = numFeature;
	u.ncol = 1;
	u.pdata = new DOUBLE[u.nline];
	for (i=0; i<numClass; i++)
	{
		CMatrix tempMatrix1;
		CMatrix tempMatrix2;
		for (j=0; j<numFeature; j++)
		{
			u.pdata[j] = this->paramU[i*numFeature+j];
		}
		u.tran(transU);
		invSigma0.mul(u, tempMatrix1);//sigma 矩阵是对称矩阵，所以不需要转置。
		for (j=0; j<numFeature; j++)
		{
			this->paramW[i*numFeature+j] = 2*tempMatrix1.pdata[j];
		}
		transU.mul(tempMatrix1, tempMatrix2);
		this->paramW0[i] = -tempMatrix2.pdata[0]+2*log(((DOUBLE)classSample[i])/((DOUBLE)totalSample));
	}
	delete[] sumX;
	delete[] sumXX;
	delete[] classSample;
	return ret;
}

/* 使用分类器进行分类，输入测试样本的文件名称，返回true表示程序执行正常，false表示程序执行错误。 */
bool CLdf::test(char* fileName)
{
	bool ret = true;
	ifstream ifs(fileName);
	int bufSize = 1024;
	char* buf=new char[bufSize];
	char* temp=new char[bufSize];
	int numClass = this->sdata.numClass;
	int numFeature = this->sdata.numFeature;
	DOUBLE* sample=new DOUBLE[numFeature];
	DOUBLE* value=new DOUBLE[numClass];
	int n=0;
	int i,j;
	int cnum;
	DOUBLE max;
	int index;
	if (NULL==this->resultTotal)
		this->resultTotal = new DOUBLE[numClass];
	if (NULL==this->resultRight)
		this->resultRight = new DOUBLE[numClass];
	for (i=0; i<numClass; i++)
	{
		this->resultTotal[i] = 0;
		this->resultRight[i] = 0;
	}
	while (ifs.good()) {
		ifs.getline(buf, bufSize);
		n=0;
		while (*(buf+n)!=',' && n<bufSize)
		{
			*(temp+n)=*(buf+n);
			n++;
		}
		*(temp+n)=0;
		cnum=this->sdata.searchClassName(temp);
		if (cnum>-1)
		{
			this->resultTotal[cnum] += 1;
			for (j=0; j<numFeature; j++)
			{
				n++;
				int n2=n;
				while (*(buf+n)!=',' && n<bufSize)
				{
					*(temp+n-n2)=*(buf+n);
					n++;
				}
				*(temp+n-n2)=0;
				*(sample+j)=atof(temp);
				if (this->sdata.maxValue[j] > this->sdata.minValue[j])
				{
					*(sample+j) = (*(sample+j) - this->sdata.minValue[j])/(this->sdata.maxValue[j] - this->sdata.minValue[j]);
				}
				else
					*(sample+j) = 0.5;
			}
			for (i=0; i<numClass; i++)
			{
				*(value+i) = this->paramW0[i];
				for (j=0; j<numFeature; j++)
				{
					*(value+i) += this->paramW[i*numFeature+j] * sample[j];
				}
			}
			max = *value;
			index = 0;
			for (i=1; i<numClass; i++)
			{
				if (*(value+i)>max)
				{
					max = *(value+i);
					index = i;
				}
			}
			if (index == cnum)
				this->resultRight[cnum] += 1;
		}
	}
	delete[] buf;
	delete[] temp;
	delete[] sample;
	delete[] value;
	return ret;
}