Date of Award

1988

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Science

First Advisor

Peter P. S. Chen

Abstract

A unified framework for knowledge base systems is proposed based on Entity-Relationship (ER) approach. Following the analysis and the specification of the real-world using Entity-Relationship approach, the knowledge base is implemented as a first-order logic system, a production system, or a frame-based system by mapping the appropriate symbolic data structures. An approach for analyzing and specifying real-world perceptions must provide appropriate semantic primitives. Therefore, a justification is provided for the semantic primitives proposed in Entity-Relationship approach by considering the fundamental issues in perception. A notation that allows Entity-Relationship approach to be used as a holistic representation is presented. Translation rules are provided for the conversion of ER-diagrams into symbolic data structures of first-order logic systems, production systems, and frame-based systems. The feasibility of using Entity-Relationship approach to support a natural language front-end of a knowledge base system is examined by analyzing the representation of surface and deep structures of a sentence in Entity-Relationship approach.

Pages

166

Share

COinS