An Effective Recommendation Framework for Personal Learning Environments Using Hybrid Recommendation Algorithms

In Personal Learning Environments, personalized recommendations are used to support the activities of learners and deliver the suitable learning resources to learners. It models the dynamic multi preferences of learners using the multidimensional attributes of resources and learner ratings. By using the data mining technology the cold start and sparsity problems are eliminated. It also increases the diversity of the recommendation list. It has two main modules namely an explicit attribute based recommender and an implicit attribute based recommender system. Based on the explicit multidimensional attributes of resources and historical ratings of the accessed resources the learner preference tree (LPT) built for each learner to express the interests of the learner. The weights of implicit or latent attributes of resources of the learner are considered as chromosomes in a genetic algorithm (GA) and this algorithm optimizes the weights according to the historical ratings in the second module. In nearest neighborhood collaborative filtering (NNCF), the recommendations are generated.