Indexing Ontological Structure Based on Neighbor Tree

Prevalent ontology search engines have facilitated the retrieval of web ontologies based on linguistic information. But the ability of exploring the structure information in web ontologies hasn´t been fully investigated, which is important for users to understand retrieved ontologies. Existing indexing technologies cannot be directly performed on ontological structure, because the lack of consensus definition of ontological structure and the difficulty of graph indexing on large-scale data. This paper proposes a novel approach of structural indexing of massive web ontologies using a “divide-and-conquer” strategy. The definition of ontological structure is given based on different kinds of semantic relations. The notion of CRSentence is introduced for structural analysis. A data structure named Neighbor Tree representing local structures is also proposed to construct the structural index of web ontologies. Experiments based on Falcons dataset show that our approach of structural indexing is feasible and promising.