Bayesian Modeling for Adaptive Hypermedia Systems


Abstract: In this paper we show how Bayesian models can be used advantageously for various adaptation tasks in open WWW-based hypermedia systems. We discuss how to model user knowledge about different topics and learning dependencies between these topics, how to make inferences for calculating the system's belief on a user's knowledge based on observations from exercises/observations. We illustrate this discussion by examples from an educational hypermedia system we developed for our course "Introduction to Java Programming".

This paper is available in Postscript-Format und als HTML-Dokument.