Research On Generation Techniques Of Deep Web Query Interface

With the development of network, more and more information is included in the network, but much more is hidden in the network deeply, called deep web. In order to mine the information of Deep Web, it is an urgency task to do some research on the data intergration of Deep Web.However, a very important aspect is the intergration of Deep Web query interface.Query interface is the only entrance of Deep Web, and it is usually represented by form. In this paper form is the main object of research, so firstly some basic imformation about form is introduced in the front of the paper, then some alogrithms of sorting are enumerated.Two tasks have been done in the paper: the first is the identification of deep web query interface.The form in the Web is researchful object in this paper, firstly some rules are used to extract the characteristics of form which constitute eigenvector, then C4.5 sort alogrithm is utilized to identify deep web query interface, in other words it means to find the entrance to hidden web. The superiority of C4.5 sorting alogrithm is validated on the Weka in experiment.After finding the entrance of deep web, the job is the classification of deep web query interface which is used to ensure the domain of query interface, such as music, business, news etc. The main object of research is also form,then the characters of text which ard expressed by VSM model are picked up. Then native bayes sorting alogrithm is used to classify deep web query interface to ensure the domain of query inerface.In this part several sorting alogrithms are compared in experiment, and the advantage of native bayes classifier is also proved.