Article ID: | iaor1991242 |
Country: | United Kingdom |
Volume: | 17 |
Start Page Number: | 553 |
End Page Number: | 560 |
Publication Date: | Jul 1990 |
Journal: | Computers and Operations Research |
Authors: | Turban Efraim, Trippi Robert R. |
Keywords: | artificial intelligence: expert systems |
Getting the rules from a domain expert is often a time consuming and expensive part of building ES. Methods of ES development which do not require the intervention of a knowledge engineer are receiving increasing attention. This paper discusses modes of automated learning possible in conventional (rule-based), statistical and neural network implemented classification-type ES, and their benefits and limitations.