Expert system for conceptual database design

Abstract

The acquisition of domain specific knowledge is one of the most important issues in the conceptual database (DB) design process. We suggest a methodology to get domain specific knowledge for conceptual DB design from analysis of business forms and dialogs with a user. And we design an expert system to support the methodology. In the system, we present domain specific knowledge using concept structures in order to clearly show the learning process of the knowledge. The procedures for generating EER (Extended Entity Relationship) conceptual model are composed of the following four steps: 1) The system analyzes one form and generates its partial concept structure 2) Whenever analysis of a new form is completed, the system integrates newly generated partial concept structure into global concept structure 3) The system checks global concept structure before generating the final results 4) From the constructed global concept structure, an EER model is automatically generated. The system includes DB design knowledge which is composed of tasks, rules and dialogs. The prototype system is developed to implement the above idea on the Windows environment using a tool UNIK.