Ontology Based Approach of Semantic Information Integration

In order to solve the problem of semantic heterogeneity in information integration, an ontology based semantic information integration (OSII) model and its logical framework are proposed. The OSII adopts the hybrid ontology approach and uses OWL (web ontology language) as the ontology language. It obtains unified views from multiple sources by building mappings between local ontologies and the global ontology. A tree-based multi-strategy ontology mapping algorithm is proposed. The algorithm is achieved by the following four steps: pre-processing, name mapping, subtree mapping and remedy mapping. The advantages of this algorithm are: mapping in the compatible datatype categories and using heuristic rules can improve mapping efficiency; both linguistic and structural similarity are used to improve the accuracy of the similarity calculation; an iterative remedy is adopted to obtain correct and complete mappings. A challenging example is used to illustrate the validity of the algorithm. The OSII is realized to effectively solve the problem of semantic heterogeneity in information integration and to implement interoperability of multiple information sources.