Web Site Visualisation as a User Navigation Aid

Shantha Jayalal, Pearl Brereton, Chris Hawksley


As e-society develops, web sites are containing increasing numbers of documents, often with complex interconnections. Existing tools are proving inadequate to enable users to navigate these web sites effectively. There are site maps to overcome this disorientation but these have limitations. In this paper we explore the use of a dynamic site map as a user navigation aid. We use an exponentially smoothed probability transition matrix for link prediction based on Markov theory and semantic clustering using lexical chains to obtain content similar pages. We introduce a prototype visualisation tool for overcoming the disorientation problem based on these principles. Applicability of link prediction & semantic clustering to other applications remains to be evaluated.