dtp.c

dTree是一个运行在WinCE上的文件管理软件。类似文件管理器,功能强大

  double tplwgt;                /* weight of tuples */
  int    attcnt;                /* number of attributes in dec. tree */
  int    nodecnt;               /* number of nodes in dec./reg. tree */
  clock_t t;                    /* timer for measurements */

  prgname = argv[0];            /* get program name for error msgs. */

  /* --- print startup/usage message --- */
  if (argc > 1) {               /* if arguments are given */
    fprintf(stderr, "%s - %s\n", argv[0], DESCRIPTION);
    fprintf(stderr, VERSION); } /* print a startup message */
  else {                        /* if no argument is given */
    printf("usage: %s [options] dtfile pdtfile "
                     "[[-d|-h hdrfile] tabfile]\n", argv[0]);
    printf("%s\n", DESCRIPTION);
    printf("%s\n", VERSION);
    printf("-m#      pruning method (default: clvl)\n");
    printf("         none: no pruning\n");
    printf("         pess: pessimistic pruning\n");
    printf("         clvl: confidence level pruning\n");
    printf("-p#      pruning parameter (default: %g)\n", param);
    printf("-q#      balance class frequencies (weight tuples)\n");
    printf("         l: lower, b: boost, s: shift tuple weights\n");
    printf("-t#      maximal height of the tree (default: no limit)\n");
    printf("-k       check largest branch "
                    "(can be very time consuming)\n");
    printf("-l#      output line length (default: no limit)\n");
    printf("-a       align values of test attributes "
                    "(default: do not align)\n");
    printf("-v       print relative frequencies (in percent)\n");
    printf("-b/f/r#  blank characters, field and record separators\n"
           "         (default: \" \\t\\r\", \" \\t\", \"\\n\")\n");
    printf("-u#      unknown value characters (default: \"?\")\n");
    printf("-n       number of tuple occurrences in last field\n");
    printf("dtfile   file containing decision/regression "
                    "tree description\n");
    printf("pdtfile  file to write pruned "
                    "decision/regression tree to\n");
    printf("-d       use default header "
                    "(field names = field numbers)\n");
    printf("-h       read table header (field names) from hdrfile\n");
    printf("hdrfile  file containing table header (field names)\n");
    printf("tabfile  table file to read "
                    "(field names in first record)\n");
    return 0;                   /* print a usage message */
  }                             /* and abort the program */

  /* --- evaluate arguments --- */
  for (i = 1; i < argc; i++) {  /* traverse arguments */
    s = argv[i];                /* get option argument */
    if (optarg) { *optarg = s; optarg = NULL; continue; }
    if ((*s == '-') && *++s) {  /* -- if argument is an option */
      while (1) {               /* traverse characters */
        switch (*s++) {         /* evaluate option */
          case 'm': optarg  = &mname;                  break;
          case 'p': param   = (float)strtod(s, &s);    break;
          case 'q': balance = (*s) ? *s++ : 0;         break;
          case 't': maxht   =   (int)strtol(s, &s, 0); break;
          case 'k': chklb   = 1;                       break;
          case 'a': dmode  |= DT_ALIGN;                break;
          case 'v': dmode  |= DT_REL;                  break;
          case 'l': maxlen  =   (int)strtol(s, &s, 0); break;
          case 'b': optarg  = &blanks;                 break;
          case 'f': optarg  = &fldseps;                break;
          case 'r': optarg  = &recseps;                break;
          case 'u': optarg  = &uvchars;                break;
          case 'n': flags  |= AS_WEIGHT;               break;
          case 'd': flags  |= AS_DFLT;                 break;
          case 'h': optarg  = &fn_hdr;                 break;
          default : error(E_OPTION, *--s);             break;
        }                       /* set option variables */
        if (!*s) break;         /* if at end of string, abort loop */
        if (optarg) { *optarg = s; optarg = NULL; break; }
      } }                       /* get option argument */
    else {                      /* if argument is no option */
      switch (k++) {            /* evaluate non-option */
        case  0: fn_dt  = s;      break;
        case  1: fn_pdt = s;      break;
        case  2: fn_tab = s;      break;
        default: error(E_ARGCNT); break;
      }                         /* note filenames */
    }
  }
  if (optarg) error(E_OPTARG);  /* check option argument */
  if ((k < 2) || (k > 3)) error(E_ARGCNT);
  if (fn_hdr && (strcmp(fn_hdr, "-") == 0))
    fn_hdr = "";                /* convert "-" to "" */
  i = (!fn_dt || !*fn_dt) ? 1 : 0;
  if  (fn_tab && !*fn_tab) i++;
  if  (fn_hdr && !*fn_hdr) i++; /* check assignments of stdin: */
  if (i > 1) error(E_STDIN);    /* stdin must not be used twice */
  if ((balance !=  0)  && (balance != 'l')
  &&  (balance != 'b') && (balance != 's'))
    error(E_BALANCE, balance);  /* check balancing mode */
  if (fn_hdr)                   /* set the header file flag */
    flags = AS_ATT | (flags & ~AS_DFLT);

  /* --- translate pruning method --- */
  if (!mname) mname = "clvl";   /* set default method name */
  method = code(pmtab, mname);  /* and get the method code */
  if (method < 0) error(E_METHOD, method);

  /* --- read decision tree --- */
  t = clock();                  /* start the timer */
  scan = sc_create(fn_dt);      /* create a scanner */
  if (!scan) error((!fn_dt || !*fn_dt) ? E_NOMEM : E_FOPEN, fn_dt);
  attset = as_create("domains", att_delete);
  if (!attset) error(E_NOMEM);  /* create an attribute set */
  fprintf(stderr, "\nreading %s ... ", sc_fname(scan));
  if ((sc_nexter(scan)   <  0)  /* start scanning (get first token) */
  ||  (as_parse(attset, scan, AT_ALL) != 0)
  ||  (as_attcnt(attset) <= 0)) /* parse attribute set */
    error(E_PARSE, sc_fname(scan));
  dtree = dt_parse(attset,scan);/* parse the decision tree */
  if (!dtree || !sc_eof(scan)) error(E_PARSE, fn_dt);
  sc_delete(scan); scan = NULL; /* delete the scanner */
  fprintf(stderr, "[%d attribute(s)/%d node(s)/%d level(s)] done ",
          as_attcnt(attset), dt_size(dtree), dt_height(dtree));
  fprintf(stderr, "[%.2fs].\n", SEC_SINCE(t));

  /* --- read table --- */
  as_chars(attset, blanks, fldseps, recseps, uvchars);
  if (k > 2) {                  /* if a table file name is given */
    table  = io_tabin(attset, fn_hdr, fn_tab, flags, "table", 2);
    if (!table) error(1);       /* read the table file */
    t = clock();                /* start the timer */
    fprintf(stderr, "reducing table ... ");
    tab_reduce(table);          /* reduce the table for speed up */
    if (balance                 /* if the balance flag is set */
    &&  (att_type(as_att(attset, dt_trgid(dtree))) == AT_SYM)) {
      tab_balance(table, dt_trgid(dtree),
                  (balance == 'l') ? -2.0F
                : (balance == 'b') ? -1.0F : 0.0F, NULL);
    }                           /* balance the class frequencies */
    fprintf(stderr,"[%d/%g tuple(s)] done ",
            tab_tplcnt(table), tab_getwgt(table, 0, INT_MAX));
    fprintf(stderr, "[%.2fs].\n", SEC_SINCE(t));
  }                             /* print a log message */

  /* --- prune decision/regression tree --- */
  t = clock();                  /* start the timer */
  fprintf(stderr, "pruning tree ... ");     /* print a log message */
  if (dt_prune(dtree, method, param, maxht, chklb, table) != 0)
    error(E_NOMEM);             /* prune the decision/regression tree */
  attcnt  = dt_attchk(dtree);   /* mark the occurring attributes */
  nodecnt = dt_size(dtree);     /* get the number of nodes, */
  maxht   = dt_height(dtree);   /* the height of the tree, */
  tplwgt  = dt_total(dtree);    /* and the tuple weight sum */
  fprintf(stderr, "[%d attribute(s)/%d node(s)/%d level(s)] done ",
          attcnt, nodecnt, maxht);
  fprintf(stderr, "[%.2fs].\n", SEC_SINCE(t));

  /* --- write decision/regression tree --- */
  t = clock();                  /* start the timer */
  if (fn_pdt && *fn_pdt)        /* if an output file name is given, */
    out = fopen(fn_pdt, "w");   /* open output file for writing */
  else {                        /* if no output file name is given, */
    out = stdout; fn_pdt = "<stdout>"; }    /* write to std. output */
  fprintf(stderr, "writing %s ... ", fn_pdt);
  if (!out) error(E_FOPEN, fn_pdt);
  if (as_desc(attset, out, AS_TITLE|AS_MARKED|AS_IVALS, maxlen) != 0)
    error(E_FWRITE, fn_pdt);    /* describe the attribute domains */
  fputc('\n', out);             /* leave one line empty */
  if (dt_desc(dtree, out, DT_TITLE|dmode, maxlen) != 0)
    error(E_FWRITE, fn_pdt);    /* describe the dec./reg. tree */
  if (maxlen <= 0) maxlen = 72; /* determine maximal line length */
  fputs("\n/*", out);           /* append additional information */
  for (k = maxlen -2; --k >= 0; ) fputc('-', out);
  fprintf(out, "\n  number of attributes: %d",   attcnt);
  fprintf(out, "\n  tree height         : %d",   maxht);
  fprintf(out, "\n  number of nodes     : %d",   nodecnt);
  fprintf(out, "\n  number of tuples    : %g\n", tplwgt);
  for (k = maxlen -2; --k >= 0; ) fputc('-', out);
  fputs("*/\n", out);           /* terminate additional information */
  if (out != stdout) {          /* if not written to standard output, */
    i = fclose(out); out = NULL;/* close the output file */
    if (i != 0) error(E_FWRITE, fn_pdt);
  }                             /* check for success and */
  fprintf(stderr, "[%d attribute(s)/%d node(s)/%d level(s)] done ",
          attcnt, nodecnt, maxht);
  fprintf(stderr, "[%.2fs].\n", SEC_SINCE(t));

  /* --- clean up --- */
  #ifndef NDEBUG
  if (table) tab_delete(table, 0);
  dt_delete(dtree, 1);          /* delete table, dec./reg. tree, */
  #endif                        /* and attribute set */
  #ifdef STORAGE
  showmem("at end of program"); /* check memory usage */
  #endif
  return 0;                     /* return 'ok' */
}  /* main() */