Managing different aspects of imperfect data in databases: a unified approach

This paper discusses varieties of imperfection aspects of information fuzziness, imprecision and uncertainty-which can be encountered together in the context of advanced and intelligent databases. Their unified treatment using a novel technique, based on fuzzy set theory is proposed. A major aim of this proposal is to be used as a basic framework to manage and incorporate imperfect data in an extended relational data model and, perhaps, in other areas of intelligent information systems. We have shown that our approach is consistent with some traditional methods in databases when dealing with these imperfection aspects individually. In query evaluation, we describe three important issues-the concepts of the upper and the lower boundary information as well as the notion of an approximation of data matching