Large ABox Reasoning with Table Schema of Relational Databases
Jin-Tak Choi, Sung‐Kook Han · 한국정보기술학회논문지 · 2012
Most of approaches for scalable reasoning over OWL ontology are to apply logic rules to the facts and axioms about individuals stored with the specialized triple format as shown in Jena. For the application of logic rules, the Description Logic Program (DLP) rules are widely used to infer ABox. In this paper we present a new large ABox reasoning method using table schema of relational databases to store the facts and axioms and DLP. The aim of this approach is to make deductions using a larger set of logic rules. The presented method can infer implicit relationships of individuals with anonymous individuals that sometimes belong to a particular class, and make it explicit by creating new individuals, if required. Further, we shift from logic paradigm of reasoning to the relational database paradigm using table schema to make some approximate deductions for individuals satisfying cardinality restrictions. The benchmark test on the reasoner implemented with above extensions bears out the enhanced reasoning capability achieved.