A Hybrid Expert System Framework: Coupling Expert System, Neural Network and Database Management System

In this paper, a new class of hybrid expert system framework using relational database systems as integrating mechanism has proposed, designed and implemented to develop expert systems and neural networks, and, to couple them to/through the database systems using client/server model and blackboard model of problemsolving. A new coupling approach (database coupling) is used to integrate expert systems and neural networks. It provides the flexibility in coupling expert systems and neural networks in different integration modes in single-user and multi-user environment. The intelligent systems are themselves stored in relational database to be concurrently accessed and shared by multiple users. The database triggers have been used to link hypotheses and to implement the control component. The resulting framework is modular, extensible, interactive and userfriendly.