Using certainty factors and possibility theory methods in a tillage selection expert system
- Resource Type
- Authors
- James A. Stone; Norman D. Clarke; Tony J. Vyn; Mary McLeish
- Source
- Expert Systems with Applications. 4:53-62
- Subject
- Computer science
business.industry
Management science
media_common.quotation_subject
General Engineering
Certainty
computer.software_genre
Expert system
Computer Science Applications
Risk analysis (engineering)
Knowledge base
Artificial Intelligence
Selection (linguistics)
Production (economics)
business
computer
Possibility theory
media_common
- Language
- ISSN
- 0957-4174
Both certainty factors and possibility theory have been successfully used to represent uncertainty in expert systems. Both of these methods were investigated during the development of a rule-based expert system for selecting tillage alternatives for corn and soybean production in Ontario. The expert system was used to test uncertainty management methods and to determine the appropriateness of possibility theory and certainty factors for representing and combining uncertainty in the knowledge base. Several techniques for representing uncertainty with these two methods are reviewed and the strengths and weaknesses of the two approaches, in terms of this project, are discussed. For expert systems in which evidence from many sources is accumulated toward a hypothesis, certainty factors appear to be more responsive than possibility theory methods. The system was validated against human experts and found to perform extremely well.