select_analysis.c

使用具有增量学习的监控式学习方法。包括几个不同的分类算法。

/* Copyright (C) 2002  Mikael Ylikoski
 * See the accompanying file "README" for the full copyright notice */
/**
 * @file
 * Select database analysis.
 * Prints all vectors in a classifier database with terms sorted in
 * order of weight. Words are taken from a dictionary file.
 *
 * Maybe also calculate overall (total sum over all classes) vector.
 *
 * Maybe also show for each word the weight for all classes:
 * hello: 0=2 1=24 2=43 3=0
 *
 * @author  Mikael Ylikoski
 * @date    2002
 */

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "dictionary.h"
#include "utility.h"
#include "vector.h"


/**
 * Add word from dictionary to array.
 *
 * @param w  word to add
 */
static void
add_word (gpointer key, gpointer value, gpointer data) {
    char **words;

    words = (char **)data;
    words[*(int *)value] = (char *)key;
}


/**
 * Main program.
 */
int
main (int argc, char *argv[]) {
    FILE *f;
    const char *str;
    char **words;	/**< Array of words from dictionary */
    int i, j, k, noc;
    dict *dt;
    vector *v;
    vv *var;

    if (argc != 3) {
	fprintf (stderr, "Usage: %s <vectorizer file> <database file>\n",
		 argv[0]);
	return 1;
    }

    /* Load dictionary */
    f = fopen (argv[1], "r");
    if (!f) {
	fprintf (stderr, "Error: Cannot open dictionary file\n");
	return 1;
    }

    dt = NULL;
    while ((str = get_next_key (f))) {
	if (!strcmp (str, "dictionary")) {
	    dt = dict_load (f);
	    break;
	}
	flush_next_value (f);
    }

    fclose (f);

    if (!dt) {
	fprintf (stderr, "Error: Cannot load dictionary\n");
	return 1;
    }

    /* Print sorted dictionary */
    j = dict_get_size (dt);
    words = my_malloc (sizeof(char *) * j);
    dict_for_each (dt, add_word);
    //for (i = 0; i < j; i++)
    //	printf ("%d %s\n", i, words[i]);

    /* Load classifier */
    f = fopen (argv[2], "r");
    if (!f) {
	fprintf (stderr, "Error: Cannot open classifier file\n");
	return 1;
    }

    /* Load global */
    noc = -1;
    while ((str = get_next_key (f))) {
	if (!strcmp (str, "noc")) {
	    str = get_next_value (f);
	    noc = atoi (str);
	    break;
	} else if (str[0] == '[') {
	    break;
	}
	flush_next_value (f);
    }

    if (noc < 1) {
	fprintf (stderr, "Error: Cannot find any classes\n");
	return 1;
    }

    while ((str = get_next_key (f))) {
	if (!strcmp (str, "[global]")) {
	    printf ("%s\n", str);
	    break;
	}
	flush_next_value (f);
    }
    if (!str) {
	fprintf (stderr, "Error: Cannot find global section\n");
	return 1;
    }

    /* Load classes */
    for (i = 0; i <= noc; i++) {
	while ((str = get_next_key (f))) {
	    if (!strcmp (str, "vec")) {
		v = vector_load (f);
		if (!v) {
		    printf ("error\n");
		}
		var = vector_sort (v);
		k = v->nel;
		for (j = 0; j < k; j++) {
		    //printf ("%d:", var[j].num);
		    printf ("%s:", words[var[j].num]);
		    printf ("%f ", var[j].val);
		}
		printf ("\n");
		continue;
	    } else if (str[0] == '[') {
		printf ("%s\n", str);
		break;
	    }
	    flush_next_value (f);
	}
    }

    return 0;
}