Designing an Intelligent Recommender System using Partial Credit Model and Bayesian Rough Set

Recommender systems have become fundamental in web-based applications and information access. They effectively prune large information spaces and provide appropriate decision making and suggestions so that users are directed toward those items that best meet their needs, preferences and interests. In web-based learning context, these systems usually neglect the learner's ability, the difficulty level of the recommended item (e.g., learning resource, exam), and the learner self-assessment. Therefore, this paper suggests an intelligent recommendation system to provide adaptive learning. The suggested system consists of two main intelligent agents First, a personalized learning resource based on partial credit model (PLR-PCM) which considers both the learner's ability and the learning resource difficulty to provide individual learning paths for learners. Second, BRS-Recommendation agent provides decision rules, as an instrument or guide for the learner's self-assessment using Bayesian Rough Set (BRS), based inductive learning algorithm. Experimental results show that the proposed system can exactly provide a learning resource closer to the learner's ability with appropriate feedback to the learner, resulting in the improvements of the learning efficiency and performance