GENERALIZATION OF THE SINGLE RULE REASONING METHOD SURE-LS FOR THE CASE OF ARBITRARY POLYGONAL SHAPED FUZZY SETS
Zsolt Csaba · 2008
Fuzzy rule interpolation (FRI) based methods present advantageous solutions for cases when the rule base does not ensure a full coverage of the input space. Several of them apply the two-step approach using in their second step a single rule reasoning (SRR) method for the determination of the conclusion from the observation and the interpolated rule. This paper presents a generalized version of the SRR method SURE-LS that was originally developed for the triangle shaped fuzzy sets. The new version extends the application area of the method to the arbitrary polygonal shaped valid membership functions. The dissimilarity of the antecedent and observation linguistic terms is measured α-cut wise and the conservation of the consequent partition’s characteristic shape type is ensured by the method of least squares.