Extracting Topics Information from Conference Web Pages Using Page Segmentation and SVM

Conference web pages display their topics information in different ways, and conferences in different domains accept papers on different topics. Automatic extraction of topics information from conference web pages is thus a difficult task and has not received much attention from the research community. In this paper, we propose a method for extracting topics information that uses a web page segmentation technique, VIPS, to segment web pages into visual blocks and uses SVM to generate extraction rules. We use conference web pages retrieved from DBWorld web site as empirical data, and experiments show that the proposed method produces satisfactory results.