Presented at: Joint Workshop on Knowledge Evolution and Ontology Dynamics (EvoDyn2011)
by Patrick Rodler, Kostyantyn Shchekotykhin, Philipp Fleiss, Gerhard Friedrich
Sequential ontology debugging aims to the efficient discrimination between diagnoses. By querying additional information the debugger can gradually reduce the number of diagnoses to be considered by the user. The selection of the best queries is of central importance for minimizing diagnosis costs. If prior fault probabilities are available, the best results are achieved by entropy based selection methods.However, given some weakly justified priors these methods bravely suggest suboptimal queries. In such a case, it is more efficient to use a no-risk method which prefers queries that eliminate 50% of diagnoses independently of any fault probabilities. However, choosing the appropriate method in advance is impossible because the quality of given priors cannot be assessed before additional information is queried.In this paper we propose a method which combines advantages of both approaches. On the one hand the method takes into account available fault probabilities and the user
Resource URI on the dog food server: http://data.semanticweb.org/workshop/evodyn/2011/paper/9
Explore this resource elsewhere: