A novel approach for Web data extraction based on XML encoding

The problem of extracting data from a Web page has been studied by many works. In this paper, we present a novel approach that extracts data records from Web pages based on techniques of XML encoding. Firstly, our approach formats a given Web data page into an XML document. Then instead of using DOM-based approaches, we make use of XML encoding model to transform the XML document into a linear sequence. Our algorithm identifies the data records of a Web page from the sequence, which avoids the complex matching between sub trees in DOM model. Moreover, we address the problem of repetitive subparts in records and propose an algorithm for data alignment. Experimental results show that our approach can extract data records accurately from web pages.