A Machine Learning Approach Classification of Deep Web Sourc

A Machine Learning Approach Classification of Deep Web Sources 1 2+ 3 4 Hexiang Xu ,Chenghong Zhang , Xiulan Hao , Yunfa Hu 1,3,4Department of Computing and Information Technology, Fudan University, Shanghai, 200433 2School of Management, Fudan University, Shanghai, 200433 +Corresponding author: chzhang@ ,xianghx01@ Abstract larger than the static pages[2], and it is still proliferating rapidly. Second, the query conditions are varying, and The classification of deep Web Sources is an important there are no rules for building the unified query interfaces. area in large-scale deep Web integration, which is still at Facing the challenge, a compromising strategy is an early stage. Many deep Web sources are structured by inevitable: Using domain-based integration[3]. Query providing structured query interfaces and results. interfaces of deep Web sources in the same domain can Classifying such structured sources into domains is one of share a lot in common. So the classification of deep Web the critical steps toward the integration of heterogeneous sources is very important in the integration. Furthermore, Web sources. To date, in terms of the Classification,