Research On Key Technologies In Ontology-based Data Integration

Because of the richness of database products, the autonomy of data model, the diversity of semantic data and some other reasons, the heterogeneous situation of data is increasingly outstanding and seriously affects the routine work. As a result, the complete extraction, comprehensive analysis and effective implement of the underlying heterogeneous data to satisfy different users with data access and sharing requirements is a severe problem in the data integration field.This thesis proposes the idea of using ontology to deal with the problems of heterogeneous data from the semantic level based on the data integration technology, designs a group of semantic mapping rules from model information to OWL ontologies and the data integration prototype system consist of the ontology construction part and query processing part etc. The main work is as follows:In the ontology construction, this thesis analyses the structural characteristics of the table-column level information, seizes the primary keys, foreign keys and other key information, and groups of semantic mapping rules from model information to OWL ontologies are proposed, and then completes the mapping from database schema information to ontology information. In the process of query creation, this thesis analyses the mapping relations between the global ontology and local ontologies, uses the hierarchical structure of ontology information, and then designs and implements the query creation algorithm based on mapping relations, by which the later steps of query processing can do work. In the process of query decomposition, this thesis analyses the field characteristics of the ontology, uses the connected feature of the central concept, and then designs and implements the query decomposition algorithm based on central concept, by which the global query is decomposed to a list of sub-queries and the combined method of results reorganization is simplified.In the end, this thesis constructs and implements the ontology-based data integration platform, on which the Query processing part and recall, precision are tested. As a result, a better unified management and sharing enquiries of heterogeneous data is achieved.