Interface schema matching with the machine learning for deep web

With the rapid development of the World Wide Web, information contained in the deep web is increasing dramatically. Since different query interfaces are heterogeneous and autonomous inherently, even in the same domain, it is a huge challenge to allow users efficiently and quickly to get their own satisfying information. Deep web query interfaces integration can solve this problem well. The interface schema matching is the foremost step in the steps of the deep web query interfaces integration. This paper takes 120 data sources as a training set and 40 data sources as a testing set. Combined with the idea of multi-strategy learning technology, a deep web interface schema matching method based on machine learning is proposed. The method transformed the schema matching problem into the machine learning classification, and achieved the schema matching automatically. In order to enhance the accuracy of the mappings, the concept of domain ontology is introduced in this paper. The experimental results show that the method has an average accuracy rate of 80%–90%.