Knowledge-Based Extrapolation of Cases: A Possibilistic Approach
Eyke Hüllermeier, Didier Dubois, Henri Prade · Studies in fuzziness and soft computing · 2002
The paper presents a formal framework of instance-based prediction in which the generalization beyond experience is founded on’the concepts of similarity and possibility. The underlying extrapolation principle is formalized by means of possibility rules, a special type of fuzzy rules. Thus, instance-based prediction can be realized as fuzzy set-based approximate reasoning. The basic model is extended by means of fuzzy set-based (linguistic) modeling techniques, including the discounting of untypical cases and the flexible handling and adequate adaptation of different similarity relations. This extension provides a convenient way of incorporating domain-specific (expert) knowledge. Our approach thus allows for combining knowledge and data in a flexible way and favors a view of instance-based reasoning according to which the user interacts closely with the system.