Masters Theses

Title

CIEGEN: a system for testing knowledge base compilation heuristics on a microcomputer

Author

Jayne D. Ward

Abstract

"The expert system has proven itself to be a valuable aid in diagnosing and treating problems in domains requiring expertise. The commercial world has been alerted to this fact and the thrust is to make the expert system portable and available on small computers. The goal of this research has been to lay the groundwork for a domain independant expert system builder on a microcomputer. The result of this effort was CIEGEN, a system consisting of a rule compiler, inference engine, and rule generator developed on the IBM PC. It is domain independant, responsible for transforming a knowledge base of rules into heuristic based decision trees, and capable of performing backward chaining consultations. The system is also heuristic independant, allowing a knowledge base to be compiled by different heuristics and compared using the log created by the inference engine. A subgoal of the development of CIEGEN has been to study the heuristics used to compile a knowledge base because the efficiency of the expert system is based on the intelligence of the heuristic. The heuristic used by EMYCIN was implemented and compared with a heuristic developed by the author. For the six types of knowledge based generated by CIEGEN's rule generator, EMYCIN 's heuristic, on the average, executed more quickly"--Abstract, leaf ii.

Advisor(s)

Gillett, Billy E.

Committee Member(s)

Dekock, Arlan R.
Kluczny, Raymond Michael

Department(s)

Computer Science

Degree Name

M.S. in Computer Science

Publisher

University of Missouri--Rolla

Publication Date

1984

Pagination

vi, 77 leaves

Note about bibliography

Includes bibliographical references (leaves 33-34).

Rights

© 1984 Jayne D. Ward, All rights reserved.

Document Type

Thesis - Citation

File Type

text

Language

English

Library of Congress Subject Headings

Heuristic algorithms
Expert systems (Computer science) -- Design
Artificial intelligence

Thesis Number

T 5114

Print OCLC #

11299429

Link to Catalog Record

Full-text not available: Request this publication directly from Missouri S&T Library or contact your local library.

http://laurel.lso.missouri.edu:80/record=b2693332~S5

This document is currently not available here.

Share

 
COinS