Abstrakt

Annotating Multiple Web Databases Using Svm

M.Yazhmozhi, M. Lavanya, Dr. N. Rajkumar

There is a far above the ground demand for deep web data search. By using search interfaces, backend database can be accessed through web. The result for the user query in search interfaces is search result record (SRR) that consists of data units. For efficient deep web data search, the SRR should be extracted out and then meaningful labels would be added to data units. This process is referred as annotation. This methodology is a time consuming process for the user and also weight values are fixed while finding the similarity. To overcome these problems, Support Vector Machine (SVM) has been used. With the help of old annotation results, the trained data has been created. This is given as input for SVM. Whenever the user gave search query, Label will be generated automatically using SVM. The SVM algorithm has been enhanced the annotation course of action.