Is there any limitation for the number of rules and the number of membership functions that we define for a particular problem? The more the number of membership functions i define, the more accurate the modeling is, with negligible amount of error. Is that a right way or is there any restriction in the number of membership functions that i define.