Ontology-Based Schema Integration Of Heterogeneous Data Sources Research

Enterprise Application Integration including business process integration, application integration, data integration, integrated standards and integration platform. But the data integration should be resolved firstly and it is very import to research and construct a stable and efficient data integration strategy, develop a corresponding data integration system for application of EAI.At the same time, the data is dispersive stored with the information of the enterprises and departments, millions of the heterogeneous data source formed in that condition. These heterogeneous data sources look like the independent information island, and the communication and operation become complexity and difficult.Every department's information island in the enterprise seems like there is no connection, but the dispersive data has the relationship actually. As the managers of the enterprise, they must face the problem of the dispersive data of theirs and analysis the relationship of the data from the high point and then make the policy of the developing of the enterprise. If each department's data could be shared, they would work more effectively, so the integration of the data becomes more important. The integration system of the heterogeneous data source could analysis the information island and find the connection of the heterogeneous and complete the integration of the information island, which could be a good way to manage and share the data.An ontology is an explicit specification of a conceptualization, which could represent the conceptualization and domain knowledge more clearly. In this paper an approach of ontology-based integration of heterogeneous data source is proposed to recognize and integrate the data in the heterogeneous data source. Firstly, every data source is described as XML documents, and then each Document Type Definition (DTD) of the XML documents is converted into a data model called DIM, finally the integrated XML document could be got through several steps, such as semantic clustering, global schema generate and so on. The method proposed in this paper is based on the electric English dictionary designed by the psychologists, linguists, computer engineers of Princeton University, which could recognize the data that contains the same or similar semantic and integrate the heterogeneous data source more accurate.