A statistical approach to incomplete information in database systems

Abstract

There are numerous situations in which a database cannot provide a precise answer to some of the questions that are posed. Sources of imprecision vary and include examples such as recording errors, incompatible scaling, and obsolete data. In many such situations, considerable prior information concerning the imprecision exists and can be exploited to provide valuable information for queries to which no exact answer can be given. The objective of this paper is to provide a framework for doing so.