Research On Ontology-based Heterogeneous Database Integration

With the rapid growth of information and Internet technologies, a huge amount of data is dispersed in many heterogeneous databases. While the "Information isolated island" problem and Deep web search needs a unified interface for retrieving the heterogeneous databases transparently. On the other hand, ontology can be used as an global concept model in database integration for its semantic reasoning ability. So, ontology-based heterogeneous database integration system is proposed as a hot research area.This paper mainly discusses two key techniques in ontology-based heterogeneous database integration system: mapping between database schema and creation of global ontology. For solving the first problem, we propose an EER-based local ontology transformation strategy, discovering the semantic of database depending on normal formula and refining ontology based on transaction queries. Also, we propose an semi-automatic method for creating global ontology, which can detect and resolve the semantic conflicts in local ontologies and thus build the mapping bridge between global ontology and local ontologies. We design a prototype to verify the improvement of our techniques.