Reinforcement learning approach towards effective content recommendation in MOOC environments

Understanding the Learner requirements is an important aspect of any learning environment as it helps to recommend the LOs in a more personalized manner. With the growing demand for MOOCs offered by coursera, edx, etc. the learner information plays a vital role in understanding the extent to which the learners can gain out of such courses. The Learning Management Systems (LMS) across the web uses the explicit (rating, performance, etc.) and implicit feedback (LOs used) obtained through interaction with the learners to derive such information. As the requirements of the learners varies with the individual's interest and learning background, a common approach for recommending LOs may not cater the needs of all the learners. To overcome this issue, this paper proposes reinforcement learning based algorithm to analyze the learner information (derived from both implicit and explicit feedback) and generate the knowledge on the learner's requirements and capabilities inside a specific learning context. The reinforcement learning system (RILS) implemented as a part of this work utilizes the knowledge thus generated in order to recommend the appropriate LOs for the learners. The results have highlighted that the knowledge derived from the learning information analysis proved to effective in generating personalized recommendation policies that can cater the context specific requirements of the learners.