sort.cpp

对给定数据集合进行分类。是数据挖掘算法的KNN的分类实例。

// Sort.cpp : Defines the entry point for the console application.
//

#include "stdafx.h"
#include "Sort.h"
#ifdef _DEBUG
#define new DEBUG_NEW
#endif

#include<fstream>
#include<vector>
#include<math.h>
#include<string>
// The one and only application object

CWinApp theApp;

using namespace std;

struct Iris
{
	double Attribute[4];//SepalLength,SepalWidth,PetalLength,PetalWidth
	string  ClassName;
};

const int K = 20;

double  IrisDistance(Iris a, Iris b)
{
	double d = 0;
	for (int i = 0; i < 4; i++)
		d += pow(a.Attribute[i] - b.Attribute[i],2);
	return d;	
}

void OutPut(int c1, int c2, int c3, Iris TestData)
{
	string result;
	if (c1 >= c2 && c1 >= c3)  
		result = "Iris-setosa";
	else if (c2 >= c1 && c2 >= c3)
		result = "Iris-versicolor";
	else
		result = "Iris-virginica";
	cout<<TestData.Attribute[0]<<'\t'<<TestData.Attribute[1]<<'\t'<<TestData.Attribute[2]<<'\t'
	<<TestData.Attribute[3]<<'\t'<<TestData.ClassName<<'\t'<<"Sort:"<<result<<'\t';
	if (result == TestData.ClassName)
		cout<<"T";
	else
		cout<<"F";
	cout<<endl;

}

void  Sort(vector<Iris> DataSet, Iris TestData)
{
	vector<int> KPos;//K个训练数据的数组下标号
	for (int i = 0; i != K ;i++)
		KPos.push_back(i);//初始化

	for (vector<int>::size_type i = K; i != DataSet.size() ; i++)//遍历其他训练数据
	{
		double m_distance = IrisDistance(DataSet[i], TestData);

		int m_MaxDistancePos = KPos[0],m_MaxDistanceKPos = 0;
		double m_MaxDistance = IrisDistance(DataSet[m_MaxDistancePos], TestData);
		for (vector<int>::size_type j = 1; j != KPos.size();j++)
		{
			if (m_MaxDistance <= IrisDistance(DataSet[KPos[j]], TestData))
			{
				m_MaxDistancePos = KPos[j];
			    m_MaxDistanceKPos = j;
			}
		}

		if (m_distance <= IrisDistance(DataSet[m_MaxDistancePos], TestData))		
			KPos[m_MaxDistanceKPos] = i;
	}

	int countSetosa = 0, countVersicolor = 0, countVirginica = 0;
	for (vector<int>::size_type  i = 0; i != KPos.size(); i++)
	{
		if (DataSet[KPos[i]].ClassName == "Iris-setosa")
			countSetosa++;
		else if (DataSet[KPos[i]].ClassName == "Iris-versicolor")
			countVersicolor++;
		else
			countVirginica++;
	}
	
	OutPut(countSetosa, countVersicolor, countVirginica, TestData);
}


int _tmain(int argc, TCHAR* argv[], TCHAR* envp[])
{
	int nRetCode = 0;

	// initialize MFC and print and error on failure
	if (!AfxWinInit(::GetModuleHandle(NULL), NULL, ::GetCommandLine(), 0))
	{
		// TODO: change error code to suit your needs
		_tprintf(_T("Fatal Error: MFC initialization failed\n"));
		nRetCode = 1;
	}
	else
	{
		// TODO: code your application's behavior here.
		
		/**********************************读入训练数据**********************************/
		ifstream infileDataSet;
		infileDataSet.open("iris.data");
		vector<Iris>  DataSet;//训练数据
		Iris  Data;
		int i = 0;//i记录对应的属性位置
		char ch, chPrior = '1';//ch为读入的字符
		double m_Attribute = 0;//将读入的字符转换为double型数据

		while(infileDataSet.get(ch))//读入字符
		{
			if (ch != '.')
			{
				if ('0' <= ch && ch <= '9')
				{
					if (chPrior != '.')
						m_Attribute = m_Attribute*10 + (ch-48);
					else
						m_Attribute += 0.1*(ch-48); 
				}
				else//, 字母 _ 换行符
				{
					if (ch == ',' )
					{
						Data.Attribute[i] = m_Attribute;
						i++;
						m_Attribute = 0;//清零
					}
					else if (ch == '\n')
					{
						i = 0;
						DataSet.push_back(Data);
						Data.ClassName.clear();
					}
					else
						Data.ClassName.append(1,ch);
				}
			}

			chPrior = ch;
		}
		infileDataSet.close();

		/**********************************读入测试数据**********************************/
		ifstream infileTestDataSet;
		infileTestDataSet.open("TestDataSet.txt");
		vector<Iris> TestDataSet;//测试数据
		Iris  TestData;

		i = 0;
		chPrior = '1';
		m_Attribute = 0;
		while(infileTestDataSet.get(ch))//读入字符
		{
			if (ch != '.')
			{
				if ('0' <= ch && ch <= '9')
				{
					if (chPrior != '.')
						m_Attribute = m_Attribute*10 + (ch-48);
					else
						m_Attribute += 0.1*(ch-48); 
				}
				else//, 字母 _ 换行符
				{
					if (ch == ',' )
					{
						TestData.Attribute[i] = m_Attribute;
						i++;
						m_Attribute =0;//清零
					}
					else if (ch == '\n')
					{
						i = 0;
						TestDataSet.push_back(TestData);
						TestData.ClassName.clear();
					}
					else
						TestData.ClassName.append(1,ch);
				}
			}
			chPrior = ch;
		}
		infileTestDataSet.close();

		cout<<"-------------------------------------Result-------------------------------------";
		for (vector<int>::size_type  k = 0 ; k != TestDataSet.size(); k++)
			Sort(DataSet, TestDataSet[k]);
	}
	return nRetCode;
}