> From:	IN%"kolarov@interval.com"  "krasimir kolarov"  3-FEB-1993 20:46:54.17
> To:	IN%"CSCON143@uoft02.utoledo.edu"  "Erik Mayer, University of Toledo"
> CC:	
> Subj:	RE: Follow-up
> 
> Erik,
> 
> I did get your last file declare.scs but the compilation 
> blew after that. In the declaration of "population" there 
> is no "initialstrength" component.
> 

	I know what happened.  There are one of two ways to fix this:

	A: Modify the poptype structure in delare.scs:

		typedef{
			 .
			 .
			 .
			 float initialstrength = 10.0;   /* add this line */
		       }poptype;		

      or B: Change the line in readclassifier() subroutine in perform.scs: 

		class->stength = population.initialstrength;

	   to:

		fscanf("%f",&class->strength);

	   then modify the data files so that each classifier has a strength
	   after it.  For example:


	       	###000:0 10
		###100:1 10
		##0#01:0 10

	   instead of:

		###000:0
		###100:1
		##0#01:0


	   All the classifiers in the data files have the initial strength of
           10.  I believe I introduced the population.initialstrength variable
	   because all of the initial strengths were the same and I think I
	   had some trouble reading in the initial strengths from the data
           file. It was a while ago when I translated this code.

> Krasimir
> Krasimir Kolarov
> Interval Research Corp., 
> 1801C Page Mill Road,
> Palo Alto, CA 94304, 
>
					- Erik Mayer 
___
  Internet: cscon143@uoft02.utoledo.edu    
  Bitnet: cscon143@uoft02.BITNET                 



