Knowledge Based System and Database Management System: An Integrative Framework

capable of powerful and efficient knowledge based processing, called Knowledge Based Management System (KBMS) has become popular as a result of merger of these two technologies. KBMS is a system providing highly effective management of large shared knowledge bases for knowledge directed systems (Brodie, 1986). A major reason for the integration of these two technologies is the realization that these are complementary technologies with the potential for enhancing the capabilities of AI systems with database features, and database management systems (DBMS) with learning and other features found in AI systems. On the database side, the ability to provide functionality in information systems using deductive, inductive and plausible reasoning, knowledge representation, heuristic search, knowledge validation and refinement provides opportunities to extend the scope and use of database management systems. On the AI side, the use of database features such as query optimization, concurrency, integrity constraints, data security, and error recovery, provide the facility to efficiently store and use the knowledge that is required for AI systems. There have been criticisms in recent times of the failure of knowledge based systems (KBS)/ expert systems delivering their promised potential and the slow growth in the use of these systems in actual practice. “Are these systems merely an intellectual exercise for the academic community with only marginal utility value for practitioners?” has been a major question facing the practitioner. Although many expert system shells and software are available in the market, users are yet to use these in a significant way in their Knowledge based management systems has gained significant importance in the last few years, primarily from the enhanced capabilities achieved through integration of the two technologies artificial intelligence and database management systems. This paper develops a framework for integrating the two technologies. The three dimensional framework links the various knowledge representation schemes with the data models used in databases and the architectures used for linking the knowledge base with the database system. This framework facilitates in communicating the latest developments in the field of knowledge based management systems and also aids the designer in making the appropriate choice from the various options available within the two technologies. The research systems in the field are mapped onto the framework. An analysis of the mapping reveals that some research areas are more popular than others. Rule-based knowledge representation with loose coupling to relational databases are found to be popular for integrating the two systems. Integrated solutions with object-oriented formalisms are also becoming common in recent years. Potential areas for future research are identified. Also, the implications of the research for the practitioner and the strategies for commercial exploitation of KBMS are discussed.