ascfoil.cpp

关联分类算法采用贪心算法发现高质量分类规则

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

#include "StdAfx.h"
#include <time.h>

#include "global.h"
#include "HashTable.h"
#include "DESCRIPTION.h"
#include "TUPLE.h"
#include "LITERAL.h"
#include "RULE.h"
#include "RULESET.h"

#include "DataSource.h"


#ifdef _DEBUG
#define new DEBUG_NEW
#undef THIS_FILE
static char THIS_FILE[] = __FILE__;
#endif

/////////////////////////////////////////////////////////////////////////////
// The one and only application object

using namespace std;

int main(int argc, const char* argv[])
{
	int nRetCode = 0;

	//FOIL

	clock_t old_tick=clock();
	double accuracy=0;
	int num_rules=0;
	int num_dataset=0;
	for(int i=0;i<10;i++)
	{
		printf("Training set #%d\n",i);
		CString suffix;
		suffix.Format("%d.txt",i);
		DataSource source;
		bool existed=source.ReadData("train"+suffix);
		if(!existed) continue;
		else num_dataset++;
		num_rules+=source.GenRulesFOIL("rule"+suffix);

		DataSource source_test;
		source_test.ReadData("test"+suffix);
		source_test.ReadRules("rule"+suffix);
		accuracy+=source_test.Classify("result"+suffix);

		if(i==0)
		{
//			source_test.IntepretRules("rule_semantics.tab","rule_interpreted.tab");
		}

	}
	int time=clock()-old_tick;
	FILE* fout=fopen("summary.txt","w");
	fprintf(fout,"\nAccuracy: %g%%\n",accuracy*100.0/num_dataset);
	fprintf(fout,"#Rules: %.1f\n",(double)num_rules/num_dataset);
	fprintf(fout,"%.3f seconds\n",time*1.0/CLOCKS_PER_SEC/num_dataset);
	fclose(fout);
	printf("\nAccuracy: %g%%\n",accuracy*100.0/num_dataset);
	printf("#Rules: %.1f\n",(double)num_rules/num_dataset);
	printf("%.3f seconds\n",time*1.0/CLOCKS_PER_SEC/num_dataset);

/*
	DataSource source;
	source.ReadData("train.txt");
	source.GenRulesFOIL("rule_foil.txt");
	source.ReadRules("rule_foil.txt");
	source.EvaluateRules("rules.txt");
*/
	return nRetCode;
}