Artificial intelligence

Abstract: In the paper the new learning algorithm of fuzzy classifier (FK) is proposed, which uses the genetic procedures to simultaneously adjust both the rule base and data base (the parameters of membership function of rule premises) of FK. The specially developed genetic procedures permit to optimize in parallel several criteria, responsible for classification accuracy, simplicity and compactness of fuzzy classifier. The comparative analysis of developed algorithm on the testing dataset Wine shows its advantage over foreign analogs according to interpretability of results preserving the high classification accuracy.

Keywords:

References: