Personalized E Learning Systems Based On Automatic Approach

To solve the problems related to the personalization in e-learning systems, namely the problems of assessing the satisfaction of learners concerning the learning devices at their disposal or problems of style detection, the satisfaction measure has always been considered as a decisive means for instructors to rectify both the learning strategies and the content associated with these strategies. However, this measure of satisfaction when it is treated in a traditional setting, namely the measurement by means of a questionnaire, suffers from certain limitations that are closely linked to the learner's consciousness which can give rise to hazardous responses that may influence the results of the questionnaire. To remedy the shortcomings of the questionnaire, the automatic methods of satisfaction measurement were introduced with the objective of processing the data resulting from the interaction of the learner with the system. In this way, classification techniques are used. Our work is staged within this framework and consists of presenting an automatic method of satisfaction measurement to evaluate the effectiveness of adaptive E Learning systems. The proposed approach is based on the analysis of the learner's traces, namely the durations associated with the different learning contents and the number of visits of these contents made by the learner. To achieve this goal, we use the space Density-based clustering algorithm of clustering applications with noise. The performances of our approach are illustrated by simulation test.