A Study Of Ontology-based Heterogeneous Data Integration On Education Resources

With the development of computer technology, traditional education modes have been affected. Increasing education resources are distributed on the Internet by the digital forms, so a great number of heterogeneous data appear accordingly. Data integration focuses on dealing with the semantic of heterogeneous data, and introducing ontology into data integration can well solve the problem.Ontology can formally describe the semantic relationship between concepts in the field. Eventually, it will be applied into the practical field for setting up shared and public semantic modes and provides preconditions for the operation on heterogeneous data. Through the study on the existing data integration technology of ontology, together with mixed ontology method, the frame of Ontology-based Heterogeneous Data Integration on Education Resources is designed and detailed research on each part is described in this dissertation.In this paper, we mainly study how to integrate the data in different relational databases, and how to build the local ontology for each data resource,which is in accordance with relational schema of the relational database. Firstly, building local ontology should extract relational schema from local data resource. Then, according to some mapping rules, the local ontology is created through the shifting from relational schema to local ontology. Referring to Education Resource Metadata Specification, the global ontology of education resources is set up on the platform of Prot ege. The semantic relations can be found between local ontologies and the global ontology. Through the mapping rules between them, integration of heterogeneous data can be realized. The mapping is built mainly based on the similarity of concept semantics, including attribute similarity, instance similarity and relation similarity. Query process is also an important part of data integration system. The global query is generated when users make a search request through searching interface; then, a proper law is needed to decompose the global query into the queries on local ontologies, which, at last, is interpreted into the query statement that can be realized by each local relational data sources.