Modeling Approach to Learner Based Ontologies for the Recommendation of Resources in an Interactive Learning Environments Mohammed Kamal Rtili PhD student, Lirosa Lab, Abdelmalek Essaâdi University, Faculty of Sciences, Tetouan, Morocco Email: rtili.kamal@gmail.com Mohamed KHALDI and Ali Dahmani Abdelmalek Essaâdi University, Faculty of Sciences, Tetouan, Morocco Email: medkhaldi@yahoo.fr, alidahmani@hotmail.com Abstract—Today, the web contains multitude sources of information and knowledge that can be used as learning materials, that’s why users are faced with a large number of irrelevant answers returned by the classical information search tools. During the last decade, recommendation systems have emerged as an effective means to reduce the complexity of information search, these recommendation systems are based on a learner’s profile. The construction of user profiles is at the center of the issues raised in the study of mechanisms for personalization or recommendation of resources to users that takes into account their specific needs. These profiles are constructed and enriched with the user interaction with the system. The objective of this paper is to present our approach for modeling a learner within our recommendation system. This is to develop a system to collect the learner’s interaction Traces and process them, in order to propose in an automatic way educational resources adapted to their needs, through a set of agents which interact with each other. Index Terms—ILE, Trace, trace model, learner profile, modeling, ontology, multi-agent systems, semantic web. I. INTRODUCTION ILE (Interactive Learning Environments) is a computing environment that uses the web as a medium for disseminating knowledge and helping the various actors interact with each other, it aims to promote, accompany and validate learning. There are several categories of an ILE, in particular the microworlds, intelligent tutoring and adaptive hypermedia. According to [1], four types of models are specified when designing an ILE, the domain model, the learner model, the pedagogical model (tutor), the expert model and the interaction model. The learner modeling is a field of research that has been the subject of several publications, it consists of all treatments allowing developing and updating relevant information about the learner from analyzing his behavior, this analysis most often consist of observing and Tracing the information of the learner activities in the system during a learning session, followed by analyzing and interpreting it. This article is organized as follows: section 2 presents the context and issues of our work, section 3 describes the theory of the Trace on which our module is based, section 4 presents the multi-agent technology, section 5 presents the general architecture of our Traces' collection module and the final section presents the conclusion and offers ideas for future developments. II. WORK CONTEXT AND PROBLEMATIC With the increasing number of websites as shown by the statistics Netcraft4 (over 644 million websites in 2013), the mass of data exchanged on the Internet is an advantage for universal access to information. In parallel, it is also a challenge because it requires significant processing to filter the data returned and find the relevant information for the user. In fact, the recommendation is a scientific field that seeks to personalize access to information for a given user, and thus facilitate his choice of content in a too extensive catalog so he can get an overall idea. In practice, recommender systems, based on knowledge about a user, filter a set of content and produce a list, often ordered, considered relevant for him. Our research is part of the customization of ILE. The goal of our research is to study and develop a recommendation system based on Traces that will enrich the services of an ILE, it is to consider the interaction Traces left by the learners as a source of knowledge which the system can exploit to propose pedagogical resources adapted to targeted learners in order to increase their satisfaction during a learning session. In fact, we have identified four steps for our system [2]: The collection of data based on the learner’s behavior, Data processing, the construction of behavioral patterns of the learners, and the recommendation of resources. In order to achieve the above-mentioned steps, we will create a system comprising of several modules, each of which has its own task (figure 1). The Traces left by the learners on the system allow to evolve this profile over time. 340 JOURNAL OF EMERGING TECHNOLOGIES IN WEB INTELLIGENCE, VOL. 6, NO. 3, AUGUST 2014 © 2014 ACADEMY PUBLISHER doi:10.4304/jetwi.6.3.340-347