Application of Set-theoretic Granular Computing

Abstract

Uncertain knowledge management is a key issue in the expert system. By analysis of the reasons for knowledge’s uncertainty in complete universe and uncompleted universe, two methods based on set-theoretic granular computing are presented to deal with the uncertainty knowledge which is the rough observable method and rough inclusion method respectively. The rough set can be approximately expressed as a union of some basic sets on the given universe using the concepts of rough membership function and rough observable set. Experiments on the information system show that the original indefinite sets can be described with its approximate sets, which have the certain knowledge. So the uncertainty sets have the certainty at some degree, and from which one can get more deterministic information.