>In article miq@worak.kaist.ac.kr, ljpark@eekaist.kaist.ac.kr (Lae-Jeong Park ) writes:
>>I'd like to know if there are papers that focusing the relationship
>>between operation principle of "classifier systems" and "fuzzy logic controller".
>>I think that the two share common features to each other.
>>Your comments and opinions are welcome.
>We think that a fuzzy classifier can be see as an incomplete fuzzy controller,
>in fact a fuzzy controler based on symbolic fuzzification without
>the defuzzification.
> <snip>
The quoted claim says classifiers are controllers. But since either
sort can be made to uniformly approximate the other, I think this
distinction is artificial and unhelpful. Someone ought to give a good
PRACTICAL reason for this distinction -- one that's useful to those
who do not care to look inside the system. I think it'll be tough to
manage.
Fred Watkins, Ph.D.