A NEW APPROACH TO INEXACT REASONING FOR RULE-BASED SYSTEMS

Abstract

This paper presents a new algorithm for dealing with inexact reasoning problems, where the certainty factors of the rules and the truth values of the conditions appearing in the rules are represented by linguistic terms. The algorithm performs inexact reasoning via repeatedly transforming an augmented linguistic truth state vector T by an augmented linguistic rule matrix F. Given the linguistic truth values of some conditions, the algorithm can perform inexact reasoning to evaluate the linguistic truth values of other conditions automatically.