Please use this identifier to cite or link to this item: http://hdl.handle.net/1959.14/156606
3 Visitors3 Hits0 Downloads
Incremental learning via exceptions for agents and humans : evaluating KR comprehensivility and usability
Pacific Rim International Conference on Artificial Intelligence (11th : 2010) (30 August - 2 September 2010 : Daegu, Korea)
Zhang, Byoung-Tak and Orgun, Mehmet A.. PRICAI 2010 : trends in artificial intelligence : 11th Pacific Rim International Conference on Artificial Intelligence, Daegu, Korea, August 30-September 2, 2010 : proceedings, p.655-661
Acquiring knowledge directly from the domain expert requires a knowledge representation and specification method that is comprehensible and feasible for the holder and creator of that knowledge. The technique, known as multiple classification ripple down rules (MCRDR), is novelly applied to the problem of building and maintaining a library of training scenarios for use by customs and immigration officer trainees in our agent-based virtual environment which may be indexed for retrieval based on the rules associated with them. Our evaluation study aims to demonstrate the utility of the MCRDR combined case and exception structure rule-based approach over standard rules alone and a non-case-based approach.