A Method of Ontology Partitioning Based on Traversing Graph
QU Hui-fan · Journal of Southwest University of Science and Technology · 2013
Due to high time complexity of the traditional methods of ontology partitioning,this paper proposed a new method based on traversing graph. Firstly,the triples( i. e. subjects,predicates and objects) were found by a query engine. Secondly,the corresponding ontology graph was constructed using RDF,where nodes correspond to subjects and objects,and edges were associated with predicates. Finally,the directed graph was depth-first traversed to achieve the partitioning. In total,the proposed method is of linear complexity O( n + e) with n being node number and e being edge number,while the traditional partitioning methods are of O( n2). It is obvious that the new method is much better.