Research on Extract the Schema of Query Interfaces

As the main approach to obtain the Deep Web data is to fill query interface provided by the pages, and then obtain them by submitting a query request to the Deep Web server, so an important step to access the Deep Web resources is to analyse the query request of Deep Web server effectively. However, the query interface is designed under different schemas and uses different language, thus it makes the extraction work of high-precision query interface schema changeable. To improve accuracy of schema extraction and to achieve interpretation of the query interfaces at semantic level, this paper proposes a new definition of query interface schema, and designs a kind of schema extraction method which based on query interface visual information and page information. The experiment adopts TEL-8 data sets of UIUC, and the experimental results show that the method of this paper has reached over 90% accuracy in different areas, in some areas even more than 95% accuracy, thus it has good feasibility and practicability.