Resolving scalability issue to ontology instance matching in Semantic Web

Abstract

Ontology instance matching is a key interoperability enabler across heterogeneous data sources in the Semantic Web and a useful maneuver in some classical data integration tasks dealing with the semantic heterogeneous assignments. Though most of the research has been conducted on ontology schema level matching so far, with the introduction of Linked Open Data (LOD) and social networks, research on ontology matching is shifting from ontology schema or concept level to instance level. Since heterogeneous sources of massive ontology instances grow sharply day-by-day, scalability has become a major research issue in ontology instance matching of semantic knowledge bases. In this paper, we propose an efficient method by grouping instances of knowledge base into several sub-groups to address the scalability issue. Then, our instance matcher, which considers the semantic specification of properties associated to instances in the matching strategy, works by comparing an instance within a classification group of one knowledge base against the instances of same sub-group of other knowledge base to achieve interoperability. A novel approach for measuring the influence of properties in the matching process is also presented. The experiment and evaluation depicts satisfactory results in terms of effectiveness and scalability over baseline methods.