User-profiling and personalisation technologies typically take an empirical approach to the analysis of a user's behaviour in order to build a model of their preferences, whether these be purchasing choices, or Web page content. A variety of techniques ranging from simple statistics to machine learning algorithms (k-nearest neighbour, naive Bayes, SVM) have been utilised to deliver personalisation solutions. Recently, the emergence of the Semantic Web as a vision for the next generation of Web technologies has provided the context for us to revisit personalisation from an analytical (knowledge-based) rather than empirical view.

In this project we are investigating how Semantic Web technologies such as RDF, ontologies, etc. can be used with machine learning methods to induce user-profiles (themselves encoded using RDF); another aim of the project is to explore how such profiles can be utilised within agent-based systems.

• Pete Edwards
• Alun Preece
• Gunnar Grimnes

• An Empirical Investigation of Learning From the Semantic Web
  In Proceedings of ECML/PKDD-2002 Semantic Web Mining Workshop, P Edwards, G A Grimnes & A D Preece.

• Granite Nights Test Platform
• SemanticWeb.org
• MLnet Information Service
• KDnet Information Service
• Semantic Web Mining Workshop 2001
• Semantic Web Mining Workshop 2002