Constructing Accurate Fuzzy Rule-Based Classification Systems Using Apriori Principles and Rule-Weighting

摘要: A fuzzy rule-based classification system (FRBCS) is one of the most popular approaches used in pattern problems. One advantage a its interpretability. However, we're faced with some challenges when generating rule-base. In high dimensional problems, we can not generate every possible rule respect to all antecedent combinations. this paper, by making use data mining concepts, propose method for generation, which result rule-base containing rules different lengths. As next phase, rule-weight as simple mechanism tune classifier and new ruleweight specification purpose. Through computer simulations on sets from UCI repository, show that proposed scheme achieves better prediction accuracy compared other non-fuzzy systems past.