dtree.h

C++编写的数据挖掘决策树算法 Mathimatics-Numerical algorithms

#ifndef _DTREE_H_
#define _DTREE_H_
#include <string>
#include <vector>
#include <set>
#include <ctime> 
#include <algorithm>
#include <cmath>
using namespace std;

int trainAllNum = 500;															//所有训练数据的范围
int testAllNum = 190;															//所有测试数据的范围
const int MaxAttr = 15;															//属性的个数
const int ArrtNum[MaxAttr] = {2, 2, 2, 4, 3, 14, 9, 2, 2, 2, 2, 2, 3, 2, 2};	//每种属性对应的取值个数

struct OriganData
{
	char A1;
	double A2;
	double A3;
	char A4;
	string A5;
	string A6;
	string A7;
	double A8;
	char A9;
	char A10;
	double A11;
	char A12;
	char A13;
	double A14;
	double A15;
	char label;
};

struct StoreData
{
	int A[15];
	char label;
};

struct TNode
{
	int attrNum;		//选取的是哪个属性
	int attr;			//属性的取值
	char label;			//例子的正负
};

struct DTree
{
	TNode node;
	vector<DTree*> childs;
};

void init();
int creatTree(DTree*& p, const vector<StoreData> &examples, vector<int> &attributes);
void readData(vector<OriganData> &data, const char* fileName);
void selectData(vector<OriganData> &data, vector<OriganData> &subdata, unsigned int selectDataNum, int dataNum);
void processConValue();
bool header(const OriganData &d1, const OriganData &d2);
double Gain(double p1, double s1, double p2, double s2);
bool allTheSame(const vector<StoreData> &samples, char);
char mostCommonAttr(const vector<StoreData> &samples);
int findBestArrt(const vector<StoreData> &examples, vector<int> &attributes);
void changeData(vector<OriganData> &Otrain, vector<StoreData> &train);
void testData();
char testTree(DTree *p, StoreData d);
void printTree(DTree *p, int depth);

#endif //_DTREE_H_