A greedy approach to rule reduction in fuzzy models

摘要: The characteristics of a fuzzy model are frequently determined by the manner in which rules constructed. Rules obtained heuristic assessment system generally linguistically interpretable and have large granularity. generation via learning algorithms that analyse training data produces precise models consisting multiple small grannularity. In this paper, greedy algorithm is presented combines rule with region merging strategy to reduce number rules. This approach differs from standard reduction techniques latter employed after base has been completed while learn-and-merge generates simultaneously expanding its applicability. objective produce both high precision.