evaluate.cpp

贝叶斯学习算法源码

	while( k < N)
	{
		MergeInKeyString(filenode,swap,k,N);
		for( i = 0;i < N ;i ++)
		{
			strcpy(filenode.keys[i].str,swap[i].str);
		}
		k = k * 2;
	}
	
	delete swap;
	return 0;
}


int CEvaluate::MergeInKeyString(FileNode &filenode, KeyWord swap[], int k, int n)
{
	
	int i,j,l1,u1,l2,u2,m;
	l1 = 0;
	m = 0;
	while(l1 + k < n)
	{
		l2 = l1 + k;
		u1 = l2 - 1;
		u2 = (l2 + k - 1 <= n - 1) ? (l2 + k - 1) :(n - 1);	

		for( i  = l1,j = l2; i <= u1 && j <= u2 ;m++)
		{
			if(strcmp(filenode.keys[i].str,filenode.keys[j].str) >= 0)
			{
				strcpy(swap[m].str,filenode.keys[i].str);
				i ++;
			}
			else
			{
				strcpy(swap[m].str,filenode.keys[j].str);
				 j++;
			}
		}

		while(i <=  u1)
		{
			strcpy(swap[m].str,filenode.keys[i].str);
			m ++;
			i ++;
		}
		while(j <= u2)
		{
			strcpy(swap[m].str,filenode.keys[j].str);
			m ++;
			j ++;
		}
		l1 = u2 + 1;	
	}
	for( i =l1;i < n; i ++,m++)
	{
		strcpy(swap[m].str,filenode.keys[i].str);
	}
	return 0;
}

int CEvaluate::CopyFeature(Feature &K, const Feature X)
{
	strcpy(K.str,X.str);
	K.gain = X.gain;
	K.NormalMail = X.NormalMail;
	K.NormalMail_Prior = X.NormalMail_Prior;
	K.UnNormalMail = X.UnNormalMail;
	K.UnNormalMail_Prior = X.UnNormalMail_Prior;
	
	return 0;
}	


int CEvaluate::ReadInfo()
{
	char s[30];
	memset(s,0,30);
	strcpy(s,Path);
	strcat(s,"KeyInfo.dat");
	ifstream inCredit(s,ios::in);
	if(!inCredit)
	{
		cerr << "Output File could not be opened."  << endl;
		exit(1);
	}
	
	for(int i = 0;i < 2 ;i ++)
	{
		inCredit >> WordReservoirInfo[i].ClassFielName >> WordReservoirInfo[i].ClassFlag >>  WordReservoirInfo[i].FileSum >> WordReservoirInfo[i].WordsSum ;
		m_SetFileNum[i] = WordReservoirInfo[i].FileSum;
	}

	return 0;
}

int CEvaluate::ReadStopList(const char *stoptext)
{
	char str[40];
	memset(str,0,40);
	int size = 500;
	stopwordArray = new StopWord[320];
	memset(stopwordArray[0].str,0,40);
	int eof,Len;
	int i = 1;
	
	FILE *readstop;
	readstop = fopen(stoptext,"r"); 
	
	if(readstop == 0)
		return -1;
	
	for(eof=fscanf(readstop,"%s",&str);eof!=EOF && eof>0;eof=fscanf(readstop,"%s",&str))
	{
		Len = strlen(str);
		str[Len] = 0;
		if(Len != 0)
		{
			if(i >= 320)
			{
				cout << "StopWord  array out run !";
				return -1;
			}
			memset(stopwordArray[i].str,0,40);
			strcpy(stopwordArray[i].str,str);
			i ++ ;
		}
		memset(str,0,40);
	} 
	fclose(readstop);
	return i - 1;
}



bool CEvaluate::Del_StopWord(const char *Word, int Num)
{
	bool flag = false;

	if(strlen(Word) < 3 )   
	{
		flag = true;
		return flag;
	}

	int low = 1,high = Num;
	int mid = 0;
	int temp;

	while(low <= high)        //快速查找法
	{
		mid = (low + high) / 2;
		temp = strcmp (Word,stopwordArray[mid].str);
		if(temp == 0)
		{
			flag = true;
			return flag;
		}
		else if(temp < 0)
			high = mid - 1;
		else 
			low = mid + 1;
	}

	return flag;
}

int CEvaluate::Seek(Item *Sour, const char *Des, int num)
{
	int flag = 0;
	int Cnt = num;
	if(Cnt == 0)
	{
		return flag;
	}
	
	for (int temp = 0; temp < num; temp ++ )
	{
		if (!strcmp(Sour[temp].str,Des))
		{
			flag = 1;
			Sour[temp].freq =  Sour[temp].freq + 1;
			return flag;
		}
	}	
	return flag;	
}

int CEvaluate::ShellSortInFreq(Item M[], int N)        //按在文章中出现的次数排列
{
	int gap = 0;
	int n = N;
	for(gap = n / 2; gap > 0; gap /= 2)
		for(int i = gap; i < n; i++)
			for(int j = i - gap; j >= 0; j -= gap)	
			{
				if(M[j].freq < M[j + gap].freq)
				{
					Item x;
					Copy(x,M[j]);
					Copy(M[j],M[j + gap]);
					Copy(M[j + gap],x);
				}
			}
			
			return 0;
}

int CEvaluate::ShellSort(Item M[], int N)       //按单词在文章升序排序
{
	int gap = 0;
	int n = N;
	for(gap = n / 2; gap > 0; gap /= 2)
		for(int i = gap; i < n; i++)
			for(int j = i - gap; j >= 0; j -= gap)	
			{
				if(strcmp(M[j].str,M[j + gap].str) > 0)
				{
					Item x;
					Copy(x,M[j]);
					Copy(M[j],M[j + gap]);
					Copy(M[j + gap],x);
				}
			}
			
			return 0;
}

int CEvaluate::FeedBack(const char *filename,int Flag)
{
	
	int i = 0;
	int k = 0;
	int Len = 0;
	int	eof = 0;
	char str[100] = "UnInit";
	
	Item *M = new Item[1000];
	if(M == 0)
		return -1;
	
	FILE *r=fopen("Cal.dat","r");       //输入文件
	if( r == 0 )
	{ 
		return -1; 
	}
	
	int tempFileNum = ++ m_FeedSetFileNum[Flag];
	
	for(eof=fscanf(r,"%s",&str);eof!=EOF && eof>0;eof=fscanf(r,"%s",&str))
	{
		Len = strlen (str);
		if(Len >39)
		{
			break;
		}
		str[Len] = 0;
		
		//增加停用词处理
		if(Len >= 3  && Seek( M,str,i)  == 0  && !Del_StopWord(str,m_StopWordNum))
		{			
			memset(M[i].str,0,40);
			strcpy(M[i].str,str);
			M[i].Sequence = tempFileNum;
			M[i].freq = 1;
			M[i].FileCnt = 1;
			M[i].density = 0.0;
			M[i].gain = 0.0;
			M[i].flag = flag;
			M[i].SetInfo = flag;
			
			i++;
		}
		memset(str,0,40);
	} 
	
	fclose(r);

	/************************************************************************/
	ShellSortInFreq(M,i);
	int Count = ((i <= KeyNum)? i: KeyNum);
	ShellSort(M,Count);
	/************************************************************************/
	

	memset(str,0,100);
	if(Flag == 0)
	{
		strcpy(str,Path);
		strcat(str,"Spam_data1.dat");
	}
	else
	{
		strcpy(str,Path);
		strcat(str,"Normal_data2.dat");
	}
	ofstream outCredit(str,ios::out | ios::ate);
	if(!outCredit)
		return -1;
	for( i = 0;i < Count;i++ )
	{			
		outCredit << M[i].str <<' ' << M[i].freq << ' ' <<M[i].density << ' ' << M[i].Sequence <<' '<< M[i].FileCnt <<' ' <<M[i].flag << '\n';
	}

	return 0;
}

int CEvaluate::Copy(Item &K, const Item M)
{
	strcpy(K.str,M.str);
	K.freq = M.freq;
	K.density = M.density;
	K.Sequence = M.Sequence;
	K.FileCnt = M.FileCnt;
	K.flag = M.flag;
	K.gain = M.gain;
	K.SetInfo = M.SetInfo;
	
	return 0;
}

int CEvaluate::mergesort(char *file,int SetFlag)
{
	Item *M = new Item[70000];
	Item *swap = new Item[70000];
	if (M == 0 || swap == 0)
		return -1;
	
	int count = 0,countCur = 0;  //变量countCur计算当前数组的有效长度
	char str[100];
	memset(str,0,100);
	strcpy(str,Path);
	strcat(str,file);
	
	ifstream inCredit(str,ios::in );
	if(!inCredit)
	{
		cerr << "Input File could not be opened"  << endl;
		exit(1);
	}
	Item Credit;
	
	while(inCredit >> Credit.str >> Credit.freq >> Credit.density 
		>> Credit.Sequence >>Credit.FileCnt >>Credit.flag)
	{
		Copy(M[countCur],Credit);
		countCur++;
	}
	
	count = countCur;
	int i ,k;
	k = 1;
	while( k < count)
	{
		merge(M,swap,k,count - 1);
		//cout << " k = " << k << endl;
		for( i = 0;i < count - 1;i ++)
		{
			Copy(M[i],swap[i]);
		}
		cout << endl;
		k = k * 2;
	}
	
	/************************************************************************/
	
	str[strlen(str) - 4] = 0;
	strcat(str,"temp.dat");
	ofstream outCredit(str,ios::out);
	if(!outCredit)
	{
		cerr << "Output File could not be opened."  << endl;
		exit(1);
	}
	
	int CN = 0;             //计算实际合并后的数据数目
	
	for( countCur = 0;countCur < count;countCur ++)
	{
		if( atof(M[countCur].str) > 0.0000001)
			M[countCur].flag = 1;
		else
			for(int temp = 1;temp <= 30;temp ++)
			{
				if(strcmp(M[countCur].str,M[countCur + temp].str) == 0  
					&& M[countCur].flag == 0 && M[countCur + temp].flag == 0  
					&& (countCur + temp < i))
				{
					M[countCur].density += M[countCur + temp].density;
					M[countCur ].freq += M[countCur + temp].freq;
					
					if(M[countCur].Sequence != M[countCur + temp].Sequence)
						M[countCur ].FileCnt ++;
					
					M[countCur + temp].flag = 1;
				}
			}
			
			if(M[countCur].flag == 0)
			{
				CN ++;
				outCredit << M[countCur].str <<' ' << M[countCur].freq << ' ' 
					<<M[countCur].density << ' ' << M[countCur].Sequence <<' '
					<< M[countCur].FileCnt <<' ' <<M[countCur].flag << '\n';
			}
	}	

	delete M;
	delete swap;
	return 0;
}

int CEvaluate::merge(Item x[], Item swap[], int k, int n)
{
	int i,j,l1,u1,l2,u2,m;
	l1 = 0;
	m = 0;
	while(l1 + k < n)
	{
		l2 = l1 + k;
		u1 = l2 - 1;
		u2 = (l2 + k - 1 <= n - 1) ? (l2 + k - 1) :(n - 1);	
		
		for( i  = l1,j = l2; i <= u1 && j <= u2 ;m++)
		{
			if(strcmp(x[i].str,x[j].str) <= 0)
			{
				Copy(swap[m],x[i]);
				i ++;
			}
			else
			{
				Copy(swap[m],x[j]);
				j++;
			}
		}
		
		while(i <=  u1)
		{
			Copy(swap[m],x[i]);
			m ++;
			i ++;
		}
		while(j <= u2)
		{
			Copy(swap[m],x[j]);
			m ++;
			j ++;
		}
		l1 = u2 + 1;	
	}
	for( i =l1;i < n; i ++,m++)
		Copy(swap[m],x[i]);
	return 0;
}


int CEvaluate::WriteInfo(char *filename)
{
	if(filename == 0)
		return -1;
	char str[100];
	memset(str,0,100);
	strcpy(str,Path);
	strcat(str,filename);
	ofstream outCredit(str,ios::out| ios::trunc);
	if(!outCredit)
	{
		cerr << "Output File could not be opened."  << endl;
		exit(1);
	}
	for(int i = 0;i < 2;i ++)
	{
		WordReservoirInfo[i].FileSum = m_FeedSetFileNum[i];
		outCredit << WordReservoirInfo[i].ClassFielName << ' ' << WordReservoirInfo[i].ClassFlag << ' '<< WordReservoirInfo[i].FileSum << ' '<< WordReservoirInfo[i].WordsSum <<'\n';		
	}

	return 0;
}

int CEvaluate::GetFileCount()
{
	return m_nFileCount;
}