Fuzzy system modeling using linear distance rules

An approach of determining significant variables for multi-input systems is proposed by using a modified fuzzy rule form, called linear distance rules (LDR). By defining an index, we can know the influences of input signals o n output and can determine significant variables from the index. We also pr opose a simple three-layered generalized neural network to realize the LDRs . Furthermore, this approach can be extended to decompose a multi-input-mul ti-output (MIMO) system into several simpler multi-input-single-output (MIS O) systems. (C) 1999 Elsevier Science B.V. All rights reserved.