Towards an optimal personalization strategy in MOOCs

Abstract

AbstractSeveral successful initiatives have resulted from the OER movement. One of them is Massive Open Online Courses (MOOCs) which is a popular learning mode as it offers an affordable and flexible way to learn. However, the evolution of the MOOCs has some challenges. One of the major problems of MOOCs is the diversity of learners and the need to personalize the content as well as the way of delivering it. The origin of this problem is the one size does not fit all. In fact, learners have different characteristics such as their learning styles, levels of knowledge, and so on. The selection of the most suitable parameters (set of complementary learners’ characteristics) to be considered in learner’s profile is not easy in the presence of a considerable number of learners in MOOCs. One reason is because a course can be attended by many learners with varied profiles from different regions of the world. This plurality of learner profiles makes it important to develop content that can meet the needs and objectives of each learner in MOOCS. Our solution to solve this problem consists of personalizing the content of MOOC for each learner. We propose a new approach which allows to optimize the selection of the personalization parameters and to apply the appropriate personalization strategy, based on a classification algorithm. The proposed approach aims to improve the retention rate and the quality of learning in MOOCs. This approach is validated by experiments which test its success when applied to many combinations of strategies and learner profiles.