Distance-Based Ranking of Negative Answers

Suggesting negative answers to users is a solution to the problem of insufficiently many answers in conjunctive query answering over description logic (DL) ontologies. Negative answers are complementary to certain answers and have explanations in the given ontology. Before being suggested to users, negative answers need to be ranked. However, there is no method for ranking negative answers by now. To fill this gap, this paper studies ranking methods that require only information on the given query and the given ontology. Three distance-based methods are proposed. Experimental results demonstrate that these methods can effectively rank negative answers to those conjunctive queries that have certain answers in the given DL ontology.