Measures of uncertainty for a distributed fully fuzzy information system

Abstract

ABSTRACTAs an important model in the field of artificial intelligence, an information system is a database that stands for relationships between objects and attributes. A distributed fully fuzzy information system is an information system with distributed fully fuzzy data. This paper investigates measures of uncertainty for a distributed fully fuzzy information system. The fuzzy -equivalence relation, induced by a fully fuzzy information system by using Gaussian kernel method, is first obtained. Then, fuzzy information structures in a distributed fully fuzzy information system is introduced. Next, Dependency between fuzzy information structures is depicted from three aspects, information distance for calculating the difference between fuzzy information structures is proposed in the same distributed fully fuzzy information system. Moreover, properties of fuzzy information structures in a distributed fully fuzzy information system are given by means of the inclusion degree. Finally, granulation measure and entropy measure of a given distributed fully fuzzy information system is proposed by means of its fuzzy information structures. These results will be very helpful for establishing a framework of granular computing and understanding the essence of uncertainty in distributed fully fuzzy information systems.