On Ontology Mapping Based on Transfer Learning

Abstract

Construction of sharable semantic space by ontology mapping is the premise and foundation of semantic inquiry and automatic composition of web service. But the present method of ontology mapping relies on syntax and structural similarity, with no formal semantic involved during the process of mapping, which makes low recall rate of mapping relations and impossibility of automated reasoning. In this paper, through extension and presentation of OWL ontology relations, an ontology mapping method based on semantic connector is presented, so the semantic relations with small granularity can be obtained through transfer learning. Furthermore, the ontology transfer learning and the construction of semantic connector is emphasized to step forward to design and develop a group of semantic arithmetics with transfer learning capacities. The research in this paper provides a new method for ontology mapping, and it can be used to create a development and coordination environment of Web software.