An integrated rule-based reduction system

Abstract

Although user-friendly and rule-based expert system shells ease the implementation of expert systems and allow for quick prototyping, their use could easily lead to ill-structured knowledge-bases. They provide no methodical way to design and maintain rule-bases (RBs). Thus, the use of these shells does not reduce the complexity of the RB design/maintenance methodology is given. The focus is on the implementation aspect of the methodology, and the RB size reduction methodology is formalized. The result is an integrated rule reduction system (IRRS) that checks for and detects incompleteness and inconsistencies in RBs and performs RB size reduction.<>