The CINET fuzzy classifier: formal background and enhancements

Abstract

This paper describes the formal background and suggests some enhancements for the fuzzy classifier, developed by Stover et al. (1996), called the continuous inferencing network (CINET), as part of the perceptor module of the prototype intelligent controller (PIC). These enhancements include, providing a mathematical foundation to the CINET fuzzy classifier seen as the cascade of a fuzzifier and a fuzzy-aggregator, extending the functionality of both the fuzzifier and the fuzzy-aggregator by incorporating a measure for randomness (called the ambiguity degree) besides a measure for vagueness (called the membership degree), and formalizing as well as simplifying the connectives used for fuzzy-aggregation.