Automatic Integration of Deep Web Query Interfaces Based on Ontology

Query interface is used to formulate queries to receive needed data from web databases in deep web. In order to access domain-specific databases, the most important step is to construct an integration interface that allows uniform access to disparate relevant sources. Therefore, this paper proposes a novel method of integrating query interfaces based on ontology in deep web. Firstly, make full use of the complex relationships which are provided by ontology, and find the semantic relationships between attributes. Then, record the matching relationships and matching frequencies into a mapping table to reserve the structure and the order of attributes. Lastly, extract these concepts from ontology as integration interface attributes which are matched with corresponding attribute labels from local interfaces, and descend order in accordance with the matching frequencies to get an integration interface. Experiment result shows that this method has dramatically improved the matching precision of query interfaces.