pair-main.c

Ripper 分类算法

/******************************************************************************
 pair-main.c - driver program for program to do paired test
******************************************************************************/


#include <stdio.h>
#include <math.h>
#include "ripper.h"

/******************************************************************************/
char *Program="pair";
char *Help_str[] = {
    "syntax: pair pred1 pred2",
    "   perform a paired test on two prediction files",
    "",
    "options are:",
    "  -v N    set verbosity",
    NULL
};

/*****************************************************************************/

int main(argc,argv)
int argc;
char *argv[];
{
    char pred1[BUFSIZ],act1[BUFSIZ],pred2[BUFSIZ],act2[BUFSIZ];
    int p1,n1,p2,n2;
    char *file1,*file2;
    FILE *fp1,*fp2;
    BOOL right1,right2;
    int wins,losses,disagrees,total;
    double pwin,se;
    int o;

    while ((o=getopt(argc,argv,"v:h"))!=EOF) {
	switch (o) {
	  case 'v':
	    set_trace_level(atoi(optarg));
	    break;
	  case 'h':
	  case '?':
	  default: 
	    give_help();
	    if (o=='h') exit(0);
	    else fatal("option not implemented");
	}
    }

    if (optind<argc) file1 = argv[optind++];
    else {
	give_help();
	fatal("prediction files must be specified");
    }
    if (optind<argc) file2 = argv[optind++];
    else {
	give_help();
	fatal("prediction files must be specified");
    }
    if (optind<argc) {
	warning("not all arguments were used: %s ...",argv[optind]);
    }

    if ((fp1=fopen(file1,"r"))==NULL) fatal("can't open %s",file1);
    if ((fp2=fopen(file2,"r"))==NULL) fatal("can't open %s",file2);

    total = wins = losses = disagrees = 0;
    while (!feof(fp1)) {
	total++;
	fscanf(fp1,"%s %d %d %s\n",pred1,&p1,&n1,act1);
	if (feof(fp2)) fatal("prediction files are of different lengths");
	fscanf(fp2,"%s %d %d %s\n",pred2,&p2,&n2,act2);
	if (strcmp(act1,act2)!=0) 
	  fatal("prediction files out of sync at line %d",total);
	right1 = (strcmp(pred1,act1)==0);
	right2 = (strcmp(pred2,act2)==0);
	if (right1!=right2) {
	    disagrees++;
	    trace(LONG) {
		printf("// %d: c1=%s c2=%s actual=%s\n",pred1,pred2,act1);
	    }
	    if (right1) wins++;
	    else losses++;
	}
    }
    if (!feof(fp2)) fatal("prediction files are of different lengths");

    printf("%s vs %s W-L-T record:\n",file1,file2);
    printf("\t%d wins %d losses %d ties\n",wins,losses,total-disagrees);
    if (disagrees) {
	pwin = ((double) wins)/disagrees;
	se = sqrt( (pwin * (1-pwin)) / (disagrees-1.0) );
	printf("\tprob(win|disagree) %.4f  +/- %.4f",pwin,se);
	if (se>0) printf("\t%.4f su>0.5\n",(pwin-0.5)/se);
	else printf("\n");
    }
}