Trading on the Edge/Trading On The Edge - CD-ROM Toolkit (Wayzata Technology)(2031)(1994).bin/pc/pc_files/venddemo/backprop/nnrexpl.c

C/C++ Source or Header | 1993-06-08 | 6KB | 180 lines

/* nnrexpl - rank the % importance of input PEs using Explain output.     */
/* assumptions: - no input data is logged in the *.nnr file               */
/*              - there are no headers in the *.nnr file                  */
/*              - the *.nnr file is NOT appended to                       */
/*  note: accurate results require that the Explain function be run on    */
/*        a file containing all training and testing data sets.           */

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <math.h>

#define MAXLINELEN 1024*5        /* longest line read from *.nnr file */
/* #define NUMINS  4 */              /* number of input PEs; avoids prompting */

main(argc, argv)
int argc;
char **argv;
{
   FILE *f1, *f2;
   char inputLine[MAXLINELEN];
   char filename[100];           /* input filename */
   int fieldCount;               /* number of fields per data case in nnr file */
   float *column;                /* pointer to dynamic memory (column totals) */
   float grandTotal;             /* total of column totals */
   float subTotal;               /* for nets with multiple output PEs */
   long tmp;
   int  numIns;                  /* inputs for this net (user specified) */
   int  numOuts;                 /* outputs for this net (fieldcount/numIns) */
   int  i, j;

/* open i/o files */
   if (argc == 2)
      strcpy (filename, argv[1]);
   else
   {
      printf ("Enter the *.nnr results filename... ");
      gets (filename);
   }

   i = strlen (filename);
   for (j = 0; filename[j] != '.' && j < i; j++);  /* locate period */
   if (j == i)
      strcat (filename, ".nnr");   /* default the extension */
   if ( (f1 = fopen (filename, "r")) == (FILE *)0)
   {
      printf ("Cannot open input file '%s', aborting.\n", filename);
      exit (1);
   }
   if ( (f2 = fopen ("nnrexpl.out", "w")) == (FILE *)0)
   {
      printf ("Cannot open output file 'nnrexpl.out', aborting.\n");
      exit (1);
   }

/* prompt for the number of input PEs in this data */
#ifdef NUMINS
   numIns = NUMINS;
#else
   numIns = 0;
   printf ("Enter the number of input PEs in this network... ");
   while (1)
   {
      gets (inputLine);
      numIns = atoi (inputLine);
      if (numIns != 0)
         break;
      printf ("Invalid entry; please re-enter the number of input PEs... ");
   }
#endif

/* find the first input record */
   fgets (inputLine, MAXLINELEN, f1);
   if (inputLine[0] == '\n')
      fgets (inputLine, MAXLINELEN, f1);

/* count the number of fields in the first record */
   fieldCount = 0;
   for (tmp = 0; inputLine[tmp] != '\0'; tmp++)
      if (inputLine[tmp] > 0x20 && inputLine[tmp+1] <= 0x20)
        fieldCount++;
   numOuts = fieldCount / numIns;
   if (numIns * numOuts != fieldCount)
   {
      printf ("Error reading data file, or incorrect number of input PEs specified. \n");
      printf ("%d inputs is not a multiple of %d fields detected in data; aborting.\n", numIns, fieldCount);
      exit (1);
   }

/* allocate memory for the accumulation and init the array */
   column = NULL;
   column = (float *)calloc (fieldCount, sizeof(float));  /* array of column outputs from 1 data case */
   if (column == NULL)
   {
      printf ("Insufficient memory to check results; aborting.");
      fclose (f1);
      exit (1);
   }
   for (i=0; i < fieldCount; i++)
      column[i] = 0.0;


/* accumulate the values */
   while (1)
   {
      tmp = 0;
      for (i=0; i != fieldCount; i++)
      {
         while (inputLine[tmp] <= 0x20)
             tmp++;
         grandTotal = 0.0;
         sscanf(&inputLine[tmp], "%f", &grandTotal);
         column[i] += fabs(grandTotal);
         while (inputLine[tmp] > 0x20)
            tmp++;
      }

/* read the next data case */
      if (fgets (inputLine, MAXLINELEN, f1) == 0)
         break;                                       /* eof */
      if (inputLine[0] == '\n')
         fgets (inputLine, MAXLINELEN, f1);

   }  /* end while */
   fclose (f1);

/* calculate the grand total */
   grandTotal = 0.0;
   for (i=0; i != fieldCount; i++)
      grandTotal += column[i];  /* total of all column totals */

/* display and log results */
/* pass 1.. rank by number of inputs */
   subTotal = 0.0;
   printf ("\nExplain Summary for %s, %d input and %d output PEs.\n", filename, numIns, numOuts);
   printf ("relative overall importance of input variables...\n");
   fprintf (f2, "relative overall importance of input variables...\n");
   for (i=0; i < fieldCount; i++)
   {
      if ( (i+1) % numOuts == 0)
      {
         subTotal += column[i];
         printf ("   input%4d %6.0f %6.1f%%\n",
                (int)(i/numOuts+1), subTotal, (subTotal/grandTotal)*100.0);
         fprintf (f2, "   input%4d %6.0f %6.1f%%\n",
                (int)(i/numOuts+1), subTotal, (subTotal/grandTotal)*100.0);
         subTotal = 0.0;
      }
      else
      {
         subTotal += column[i];
      }
   }
/* pass 2.. if multiple outputs, rank by number of outputs */
   if (numOuts > 1)
   {
      printf ("relative importance of inputs per output PE...\n");
      fprintf (f2, "relative importance of inputs per output PE...\n");
      for (i=0; i < numOuts; i++)
      {
         printf ("output%4d\n", i+1);
         fprintf (f2, "output%4d\n", i+1);
         grandTotal = 0.0;
         for (j=0; j < numIns; j++)
            grandTotal += column[i+j*numOuts];
         for (j=0; j < numIns; j++)
         {
            printf ("   input%4d %6.0f %6.1f%%\n",
              j+1, column[i+j*numOuts], (column[i+j*numOuts]/grandTotal)*100.0);
            fprintf (f2, "   input%4d %6.0f %6.1f%%\n",
              j+1, column[i+j*numOuts], (column[i+j*numOuts]/grandTotal)*100.0);
         }
      }
   }
   fclose (f2);
   printf ("Output file is nnrexpl.out.");
   free (column);
   exit (0);
}