A partitioning algorithm for large scale ontologies

As the need and usage of semantic web grows, the number of semantic web data made up of ontology also increases. The ontology constructed for real world domains like medicine, life science and e-commerce are large in size ranging from 1000 to more than 20000 concepts. Various semantic operations like query answering, data sharing, data matching, data reuse and data integration become complicated as the size of ontology increases. Partitioning the ontology is the key solution to handle this scalability issue. This paper presents an efficient neighbour based bottom up partitioning algorithm (Refined AHSCAN) to tackle this scalability issue. The proposed partition algorithm divides the large unmanageable ontology into small manageable sub ontologies (partitions). The efficiency in the partition algorithm is achieved by reducing the computation needed for finding the neighbour similarity of the concepts and merging the partitions in bottom up hierarchy without compromising on the effectiveness (quality) of the result.