Research On Multiple Sources Schema Matching Method In The Deep Web

With the prevalence of deep Web, integrating the large amount of information hidden in the deep Web has become the focus of data integration area. Integrating the deep Web needs matching the query interfaces of the deep Web firstly. So as the key operation in deep Web data integration, multiple sources schema matching method becomes the focus of schema matching area. Based on the synthetical analyzation of the current situation of multiple sources schema matching method, the paper studies the multiple sources schema matching method from a completely new perspective. The researchment in this paper is as follows.Firstly, the MetaQuerier system and the concept of the multiple sources schema matching method are introduced, the similarities and differences of multiple sources schema matching method and general schema matching method are analysed, and some concepts which are related to the model construction and complex matching are introduced. All of the above make a foundation of the paper.Secondly, on the base of the current multiple sources schema matching methods, an improved method based on prime number is proposed, which makes up for the shortage of the current methods.The concerte algorithm and the corresponding time complexity are proposed. The algorithm is explained by the concrete examples.Thirdly, in view of the characteristics of deep Web sources and the matching results, the paper proposes a multiple sources schema matching method based on knowledge. It creates a matching knowledge database for each domain of deep Web, and it shortens the matching time by taking advantage of the matching knowledge during the matching process. The paper also proposes the concret algorithm and its time complexity.Finally, the above algorithms are validated by experiments in sources of four deep Web domains, and the experiment results are analyzed.