A Multi-Dimensional Recommendation Framework for Learning Material by Naive Bayes Classifier

Personal Learning Environment (PLE) solutions can empower learners to design ICT environments for their activities in different learning contexts. Recommender systems have been used for supporting learners in PLE-based activities. Since, in the current recommendation approaches, multidimensional attributes of resource and dynamic interests and multi-preference of learners are not fully considered simultaneously, this paper proposes a novel resource recommendation framework in order to personalize learning environments. Learner Tree (LT) is introduced to take into account the multidimensional attributes of resources and learnersʹ rating matrix simultaneously. In addition, a forgetting function also is used to reflect dynamic preference of a learner and a Bayesian classifier is used to predict rate of unrated resources. The main contribution of this paper is proposing a multidimensional data model to consider multi-preference of learner and using naive Bayes classifier to improve the quality of recommendation in the terms of precision, recall and also intra-list similarity. In addition, the proposed approach tries to satisfy the learner’s real learning preference accurately according to the real-time up dated contextual information.