Hidden Schema Extraction in Web Documents

One of the main limitation when accessing the web is the lack of explicit schema about the logical organization of web pages/sites, whose presence may help in understanding data semantics. Here, an approach to extract a logical schema from web pages based on HTML source code analysis is presented. We define a set of primary tags actually used to give a structural/logical backbone to the page. Primary tags are used to divide the page into collections, which represent distinct structural page sections; these are finally mapped into logical sections according to their semantics, providing a logical page schema. The structuring methodology is applied to some real web pages to test the approach.