Journal of Network and Computer Applications 31 (2008) 603–627 User Profile Modeling in the context of web-based learning management systems Yiouli Kritikou a , Panagiotis Demestichas a,Ã , Evgenia Adamopoulou b , Konstantinos Demestichas b , Michael Theologou b , Maria Paradia c a Department of Digital Systems, University of Piraeus, 80, Karaoli and Dimitriou Street, P.C.: 18532, Piraeus, Greece b National Technical University of Athens, 9, Iroon Polytexniou, P.C. 15780, Zografou, Greece c Pedagogical Institute, 316, Mesogion, P.C. 15341, Agia Paraskevi, Greece Received 21 November 2006; received in revised form 5 November 2007; accepted 13 November 2007 Abstract Over the past two decades, great research efforts have been made towards the personalization of e- learning platforms. This feature increases remarkably the quality of the provided learning services, since the users’ special needs and capabilities are respected. The idea of predicting the users’ preferences and adapting the e-learning platform accordingly is the focal point of this paper. In particular, this paper starts with the main requirements of an advanced e-learning system, explains the way a user navigates in such a system, presents the architecture of a corresponding e-learning system and describes its main components. Research is focused on the User Model component, its role in the e-learning system and the parameters that comprise it. In this context, Bayesian Networks are used as a tool for the encoding, learning and reasoning of probabilistic relationships, with the aim to effectively predict user preferences. In support of this vision, four different scenarios are presented, in order to test the way Bayesian Networks apply in the e-learning field. r 2007 Elsevier Ltd. All rights reserved. Keywords: Bayesian Networks; E-learning system; User modeling; User Profile ARTICLE IN PRESS www.elsevier.com/locate/jnca 1084-8045/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.jnca.2007.11.006 Ã Corresponding author. E-mail address: pdemest@unipi.gr (P. Demestichas).