Management of Latent Learning Needs in Adaptive e-Learning Systems

The construction of personalized courses taking into account needs and preferences of each learner is one of the most studied topics within the field of adaptive learning systems. Such systems assume that learning needs are well know by the teacher or, in self directed learning settings, that learners can easily determine and express own needs. Nevertheless, in real situations, learning needs often remain latent and, for this reason, unsatisfied. To overcome this limitation we propose in this paper a methodology and a prototype system enabling the elicitation of latent learning needs, as well as the automatic generation of learning experiences capable of satisfying such needs. The proposed methodology applies principles of recommending systems and also relies on a semantic representation of topics to be taught. The encouraging experimental results obtained with University students are also discussed.