Accommodating Imprecision in Database Systems: Issues and Solutions

Most database systems are designed under assumptions of precision of both the data stored in their databases, and the requests to retrieve data. In reality, however, these assumptions are often invalid, and in recent years considerable attention has been given to issues of imprecision i database systems. In this paper we review the major solutions for accommodating imprecision, and we describe issues that have yet to addressed, offering possible research directions. 1 In t roduct ion Information stored in a database is precise if it is assured to be identical to the "real world" information which it represents. When precise information is unavailable, it is often the case that some relevant information is nonetheless available. In these cases, it may be advantageous to design methods by which this information, termed imprecise information, can be stored, manipulated and retrieved. Imprecision may also be present in requests to retrieve data, when users, either intentionally or by necessity, formulate their queries in imprecise terms. Depending on the data model used, the information stored in a database may take different forms, and imprecision could affect each and every one of them. For example, in the entity