A Relevance-Directed Algorithm for Finding Justifications of DL Entailments
- Cite this paper as:
- Ji Q., Qi G., Haase P. (2009) A Relevance-Directed Algorithm for Finding Justifications of DL Entailments. In: Gómez-Pérez A., Yu Y., Ding Y. (eds) The Semantic Web. ASWC 2009. Lecture Notes in Computer Science, vol 5926. Springer, Berlin, Heidelberg
Finding the justifications of an entailment, i.e. minimal sets of axioms responsible for the entailment, is an important problem in ontology engineering and thus has become a key reasoning task for Description Logic-based ontologies. Although practical techniques to find all possible justifications exist, efficiency is still a problem. Furthermore, in the worst case the number of justifications for a subsumption entailment is exponential in the size of the ontology. Therefore, it is not always desirable to compute all justifications. In this paper, we propose a novel black-box algorithm that iteratively constructs a set of justifications of an entailment using a relevance-based selection function. Each justification returned by our algorithm is attached with a weight denoting its relevance degree w.r.t. the entailment. Finally, we implement the algorithm and present evaluation results over real-life ontologies that show the benefits of the selection function.
Unable to display preview. Download preview PDF.