Fuzzy sets of rules for system identification

摘要: The synthesis of fuzzy systems involves the identification a structure and its specialization by means parameter optimization. In doing this, symbolic approaches which encode information in form high-level rules allow further manipulation system to minimize complexity, possibly implementation cost, while all-parametric methodologies often achieve better approximation performance. this paper, we rely on concept set tackle rule induction problem at an intermediate level. An online adaptive algorithm is developed almost surely learns extent inclusion significantly contributes reproduction target behavior. Then, resulting can be defuzzified give conventional with similar Comparisons low-level show that approach retains most positive features both.