Answering Non-Standard Queries in Distributed Knowledge-Based Systems

Abstract

In this paper we present a query answering system for solving nonstandard queries in a distributed knowledge based system (DKBS). Our system is different from solving queries on a conventional distributed database or cooperative database in the sense that it discovers rules, if needed, and uses them to resolve unknown attributes. In [12], the rules used to resolve unknown attributes are discovered directly from the tables (relational databases) either locally or on remote sites. In this paper, the rule discovery process is dependent on descriptions of objects which will never be stored in our system (they either do not exist or we have no interest in storing them). Such descriptions are called either locally-negative (1-negative) or globally-negative (g-negative) terms. L-negative terms refer to the situation when only a local site of DKBS is taken into cosideration. If any site of DKBS is considered for storing the data, we use g-negative terms instead.