DocumentCode :
3310153
Title :
A new perspective on conflict resolution in market forecasting
Author :
Fishman, Mark B. ; Barr, Dean S. ; Heavner, Ellen
Author_Institution :
Eckerd Coll., St. Petersburg, FL, USA
fYear :
1991
fDate :
9-11 Oct 1991
Firstpage :
97
Lastpage :
102
Abstract :
What has been most problematical in the application of expert systems technology to market forecasting, as well as a significant theoretical issue with respect to the relative `insensitivity´ of the control structure of the expert system shell to global changes in the nature of the dynamism of the predictive domain has been that the market is sometimes predictable by one mechanism, sometimes by another, and sometimes, for practical purposes, not at all. The authors demonstrate a novel technique of conflict resolution implemented in a shell that permits changes in global market parameters to be reflected immediately in the disposition of every rule in the expert system to fire, relative to every other rule. The conflict resolution preference mechanism itself changes, in other words with the trending nature of the market, and the authors have found this to produce a very effective, market-sensitive model with substantial success (upwards of 86%) in timing buys and sells under a wide range of altogether divergent market conditions
Keywords :
expert systems; forecasting theory; stock markets; conflict resolution preference mechanism; control structure; divergent market conditions; expert system shell; expert systems technology; global changes; global market parameters; market forecasting; market-sensitive model; predictive domain; theoretical issue; trending nature; Control systems; Cost function; Economic forecasting; Educational institutions; Expert systems; Globalization; Safety; System testing; Technology forecasting; Uncertainty;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Artificial Intelligence Applications on Wall Street, 1991. Proceedings., First International Conference on
Conference_Location :
New York, NY
Print_ISBN :
0-8186-2240-7
Type :
conf
DOI :
10.1109/AIAWS.1991.236566
Filename :
236566
Link To Document :
بازگشت