Fuzzy cluster analysis for multi-antecedent rule base restructuring based on S-implication

A new fuzzy cluster analysis approach is proposed for a restructuring of multi-antecedent rule bases where rules are based on S-implication. The algorithms can be applied with a compositional tolerance relation or with an analogical tolerance relation. The adoption of a specific algorithm therefore corresponds to the inference mode to be executed on the knowledge base. It is shown that the analogical tolerance relation is highly efficient in rule base restructuring and rule firing under a proposed search schema. Finally, further studies are discussed.<>