Neighborhood-based information systems

Abstract

The concept of neighborhood systems originated from studies in topology. In this study, semantical interpretations of neighborhood systems are given by introducing certain assumptions on neighborhoods. Based upon the notion of neighborhood systems, a structure called n-closure is proposed, which is a convenient and powerful tool to model the nearness among elements in a universe. A binary relation is defined on the n-closure, based on which it is possible to rank all elements in a universe according to their nearness to a given element. The notion of information systems is enriched by adding neighborhood systems on attribute values. A Pawlak information system, (see document) In a Pawlak information system, the retrieval method is based upon an exact matching mechanism by using the equality relation =. In this study, a retrieval model is proposed in the framework of neighborhood-based information systems. Queries are relaxed by using the notion of neighborhood systems. The retrieved results are ranked by adopting the proposed ranking structure. This model provides more useful results than that of Pawlak information systems.