Optimising Ontology Classification

Abstract

Abstract. Ontology classification-the computation of subsumption hierarchies for classes and properties-is one of the most important tasks for OWL reasoners. Based on the algorithm by Shearer and Horrocks [9], we present a new classification procedure that addresses several open issues of the original algorithm, and that uses several novel optimisations in order to achieve superior performance. We also consider the classification of (object and data) properties. We show that algorithms commonly used to implement that task are incomplete even for relatively weak ontology languages. Furthermore, we show how to reduce the property classification problem into a standard (class) classification problem, which allows reasoners to classify properties using our optimised procedure. We have implemented our algorithms in the OWL HermiT reasoner, and we present the results of a performance evaluation.

Answering SPARQL Queries over Databases under OWL 2 QL Entailment Regime

“…We compared Ontop with two other systems, OWL-BGP r123 [19] and Pellet 2.3.1 [31] (Stardog and OWLIM are incomplete for the OWL 2 QL entailment regime). OWL-BGP requires an OWL 2 reasoner as a backend; as in [19], we employed HermiT 1.3.8 [14] and Pellet 2.3.1. The hardware was an HP Proliant Linux server with 144 cores @3.47GHz, 106GB of RAM and a 1TB 15k RPM HD.…”

Practical Reasoning for Defeasible Description Logics

“…However, in prac- tice, with the development of an enhanced traversal algorithm [13] and subsequent optimisations [179,77], in addition to novel techniques [76], one can vastly reduce this number. Current implementations can classify a KB on demand even if its size is in the order of hundreds of thousands of concept names [115,106].…”

Ontological knowledge bases productivity optimization through the use of reasoner combination

“…The study has analyzed the works on the optimization of ontological knowledge base performance [4][5][6][7][8][9][10].…”