method.h

利用贝叶斯分类原理实现多义词的消歧。首先利用训练语料进行训练

// Method.h: interface for the CMethod class.
//
//////////////////////////////////////////////////////////////////////

#if !defined(AFX_METHOD_H__6764876F_D132_4DE4_9C0D_5B411FAEDF43__INCLUDED_)
#define AFX_METHOD_H__6764876F_D132_4DE4_9C0D_5B411FAEDF43__INCLUDED_

#define STL_USING_ALL
#include "Stl.h"

#if _MSC_VER > 1000
#pragma once
#endif // _MSC_VER > 1000

class CContext//用上下文出现的词汇建树
{
public:
	int ContextWordID;
	int ContextWordCount;
	CContext * ChildL;
	CContext * ChildR;
};
typedef CContext * BTREE;

class wordPVS
{
public:
	int IDw;
	float Pvs;
};

class Word
{
public:
	string wordName;
//	int senseID;
	string senseID;
	friend int operator==(const Word &Left,const Word &Right)
	{
		if(Left.wordName==Right.wordName)
			return Left.senseID==Right.senseID;
		else return Left.wordName==Right.wordName;
	}
	friend int operator>(const Word &Left,const Word &Right)
	{
		if(Left.wordName==Right.wordName)
		{
			return Left.senseID>Right.senseID;
		}
		else 
			return Left.wordName>Right.wordName;
	}
	friend operator<(const Word &Left,const Word &Right)
	{
		if(Left.wordName==Right.wordName)
		{
			return Left.senseID<Right.senseID;
		}
		else 
			return Left.wordName<Right.wordName;
	}
};

class CWord//表中的元素
{
public:
	int wordID;//senseID;
	int wordCount;
	Word word;

    BTREE Context_T; 

	friend int operator<(const CWord& Left,const CWord& Right)
	{
		return Left.word<Right.word;
	}

	float Psk;
	float P_LAM;  //记入一个平滑参数
};

class CMethod  //训练参数的方法
{
public:
	void IntravTree(fstream & fout, BTREE & T, int k);
	void CountParaII();
	void CountPara();
	void CreateTree(BTREE &T, int ID);
	void SearchWord(Word woTmp);
	void SplitLine(string);
	void ReadSentence();
	void ReadSenseNo();
	CMethod();
	virtual ~CMethod();
private:
	CWord WordList;              //存每个词的信息（包括词名，ID号，语义号，词频，上下文）
	vector<CWord> vectWordList;  //保存词表（所有词的信息）
	string strTmp;
	vector<Word> vectword;       //保存上下文出现词汇及其语义ID号
	typedef vector<CWord>::iterator PPP;
	vector<PPP> vectMid;         //保存向量的元素位置
	vector<int> vectWorID;       //保存上下文出现词汇的ID号

	
};

#endif // !defined(AFX_METHOD_H__6764876F_D132_4DE4_9C0D_5B411FAEDF43__INCLUDED_)