Ontology-oriented case-based reasoning (CBR) approach for trainings adaptive delivery DOUNIA MANSOURI (1) , ABOUBEKEUR HAMDI-CHERIF (2),(3) (1),(2) Ferhat Abbas Setif University (UFAS) Faculty of Science - Computer Science Department 19000 Setif ALGERIA (1) dounia.mansouri@yahoo.fr (3) Qassim University Computer College - Computer Science Department PO Box 6688 – 51452 Buraydah SAUDI ARABIA (2),(3) shrief@qu.edu.sa Abstract: - We propose an approach that adaptively provides the reuse of previous experience of trainings contents to be used by different audiences. The representation of the main building-blocks, or learning objects, that are at the basis of these trainings, is modeled using ontologies. The approach relies on case-based reasoning (CBR) since the trainings adaptation is based on the traces left by previous learning processes. Knowledge is stored in the form of cases, rather than rules. When a new situation is encountered, the CBR system reviews the cases in an attempt to find a match for this particular training. If a match is found, then that specific case can be used to solve the new problem, otherwise it is stored as a new independent problem with a chosen default solution, introduced by the human expert. Following these lines, we develop an adaptation algorithm responsible for the required corrective actions in trainings adaptive delivery destined to diversified learners. Key-Words: - Training process, Case-based reasoning, Ontologies, Learning objects, Experience sharing, Trainings adaptive delivery 1 Introduction Our work represents a contribution to ontology- based indexing in the case-based reasoning (CBR) process. From the cognitive point of view, CBR is a general methodology that concerns the acquisition, representation and use of ad hoc experiences for addressing evolvable and novel situations. The application area is trainings adaptive delivery, a common and challenging issue. In order to ease the adaptation of trainings to different audiences, it is necessary to choose some basic configurations that are as flexible as possible to meet different environments while highlighting some local adjustments to be achieved; hence the use of ontology for indexing the cases. Moreover, in the field of the e-learning, the applications of the adaptable approach are particularly important as the e-learning process often require such adjustments on the case-by-case basis, depending on the learners’ achievements, and on a dynamic monitoring of their success in the chosen trainings. A dynamic content adaptation to the profile of a successful learner is therefore an essential feature. In this context, we choose the experience reuse approach to assist the users whereby the system memorizes and interprets the current tasks signatures, i.e. the traces left from experience (Heraud, 2002). Because heavily-relying on experience, we show that CBR offers an acceptable paradigm for addressing the issue of trainings adaptive delivery. On the basis on the arguments expanded above, we propose a model of reasoning in trainings adaptive delivery that is rooted in CBR, since the model adapts to the new learners the footsteps of similar learners who have achieved the same target profile starting from closely-related or similar profiles (Mansouri and Mille, 2007). To achieve the proposed objectives, we represent trainings in the form of ontologies used to semantically index learning objects, on the basis of the current standards of the e-learning such as SCORM and LOM. Standards are used in order to allow the homogeneity of representations of the learning Recent Researches in Computer Science ISBN: 978-1-61804-019-0 328