USING OBJECT CONCEPTS TO MATCH ARTIFICIAL-INTELLIGENCE TECHNIQUES TO PROBLEM TYPES

Using object-concepts as a matching framework, we provide guidelines f or identifying what types of problems are best served by which knowled ge-representation technique. We find that production rules are best fo r hierarchical classification problems, because they support classific ation/instantiation of data. Frames are best for data retrieval and in ference problems, because, using data abstraction, frames can operate on data within a frame. Finally, semantic networks are best for conseq uence finding problems, because of independence of the primitives in t he hierarchy. Providing guidelines for this matching is important, bec ause the success of different information systems designs have been sh own to depend explicitly on problem characteristics. (C) 1998 Elsevier Science B.V. All rights reserved.