PERSO: A system to build dynamically personalized courses in an e-learning environment (ITHET 2003) Henda CHORFI, Mohamed JEMNI ESSTT, department dInformatique 5, Av. Taha Hussein, B.P.56, Bab Mnara henda.chorfi@esstt.rnu.tn Mohamed.jemni@fst.rnu.tn Esma AIMEUR Departement d’informatique et de recherche operationnelle, Pavillon Andre Aisenstadt, C.P. 6128, Montreal H3C 3J7 (QC), Canada aimeur@IRO.UMontreal.CA Abstract - The aim of this work is to design an intelligent e-Learning Environment. We propose a system that initializes the student model with the affective and the cognitive information, which constitutes his background. When integrated in an e- learning platform, PERSO, offers to the student an individualized training program. The system is based on two approaches: Latent Semantic Analysis (LSA) and Case Based Reasoning (CBR). The first one is used to represent the curriculum and to analyze semantically the student responses. The second one is used to propose an individual training program to each student. I. INTRODUCTION With the spread of the New Technologies of Information and Communication (NTIC), we live a revolution in many domains particularly in education and training with the development of a new mode of education: the e-learning [2, 3]. E-learning, the electronic delivery of information, communication, education and training, provides a new set of tools that can add value to all the traditional learning modes-classroom experiences, textbook study, CD-ROM and traditional Computer Based Training. E-learning is a revolutionary way to empower a workforce with the skills and knowledge it needs to turn change to an advantage. It will not replace the classroom setting, but enhance it, taking advantage of new content and delivery technologies to enable learning. However, in practice, and in the majority of cases, the change affects only the way how to deliver information (electronically) but not what to deliver. The most of web based training programs are just on-line documents. Trainers create electronic versions of traditional printed student manuals, articles and reference guides then diffuse them via dedicated systems called platforms. Platforms are Web based systems that integrate services for virtual learning, teaching and informing including content creation, communication tools and user administration. Most of e-learning platforms are focused on the transfer of knowledge, and this is where their shortcomings are evident. Knowledge sharing is very important, but it does not generate the behaviour changes needed to develop competent learners. For these reasons, most of nowadays platforms are unable to detect the weaknesses of students, difficulties they meet and their needs. Furthermore, they are unable to react with them and to propose solutions to fix their problems. In this context, the objective of this work is to grief intelligence in e-learning environments to make them more efficient. One of .the main properties of an intelligent e-learning system, in our sense, is that it can built dynamically a suitable training program to each student corresponding to his preferences and his background. In this paper, we start by providing an overview and a diagnosis of an e-learning experience we have recently done in the University of Tunis. Next, we give a presentation of the two main approaches used in the system PERSO: Latent Semantic Analysis (LSA) and Case Based Reasoning (CBR). Then we present the architecture of the system.