Towards building logical views of websites

Information presented in a Website is usually organized into certain logical structure that is intuitive to users. It would be useful to model websites with such logical structure so that extraction of Web data from these sites can be performed in a simple and efficient manner. However, the recognition and reconstruction of such logical structure by software agent is not straightforward due to the complex hyperlink structure among webpages and the HTML formatting within each webpage. In this paper, we propose the Wiccap Data Model, an XML data model that maps websites from their physical structure into commonly perceived logical views. To enable easy and rapid creation of such data models, we have implemented a visual tool, called the Mapping Wizard, to facilitate and automate the process of producing Wiccap Data Models. Using the tool, the time required to construct a logical data model for a given Website is significantly reduced.