myk-means.cpp

一个基于类的k-means算法的实现。学习这个算法的一个好例子

#include "myk-means.h"
#include <fstream>
#include <cmath>
using namespace std;

void kmeans::initial(string fileName,int clusterNum)
{
	getData(fileName);
	if(data.size()<clusterNum)
		numOfCluster = data.size();
	else
		numOfCluster = clusterNum;
	center.resize(clusterNum,-1);
	cluster.resize(clusterNum);
}

void kmeans::getData(string fileName)
{
	ifstream in;
	in.open(fileName.c_str(),ios::in);
	if(!in.is_open())
	{
		cout<<"open file failed!";
		exit(0);
	}
	int size;
	in>>size;
	for(int i=0;i<size;i++)
	{
		double temp;
		in>>temp;
		data.push_back(temp);
	}
	in.close();
}

void kmeans::distribute()
{
	for(int i=0;i<numOfCluster;i++)
		cluster[i].clear();
	for(int i=0;i<data.size();i++)
	{
		int location = findCloset(data[i]);
		cluster[location].insert(data[i]);
	}
}

int kmeans::findCloset(double t)
{
	int location=0;
	double delta=9.9e+20;

	for(int i=0;i<numOfCluster;i++)
	{
		if(fabs(center[i]-t)<delta)
		{
			delta = fabs(center[i]-t);
			location = i;
		}
	}
	return location;
}

bool kmeans::findNewCenter()
{
	bool again = false;
	for(int i=0;i<numOfCluster;i++)
	{
		double avg=0;
		int size = cluster[i].size();
		set<double>::iterator ite = cluster[i].begin();
		while(ite!=cluster[i].end())
		{
			avg = avg + (*ite)/size;
			ite++;
		}
		if(avg!=center[i])
		{
			center[i]=avg;
			again=true;
		}
	}
	return again;
}

void kmeans::print()
{
	ofstream cout("mean.txt",ios::out);
	for(int i=0;i<numOfCluster;i++)
	{
		cout<<"cluster  "<<i+1<<':'<<endl;
		cout<<"center  "<<center[i]<<endl;
		set<double>::iterator ite = cluster[i].begin();
		while(ite!=cluster[i].end())
		{
			cout<<(*ite)<<' ';
			ite++;
		}
		cout<<endl<<endl<<endl;
	}
	//cout.close();
}