Heterogeneous educational resource recommender system based on user preferences

Abstract

Traditional online learning systems are based on different filtering techniques which usually rely on user behaviour towards different resources. The recommendation of resources is extracted from other users with similar behaviour. These systems usually recommend unsatisfactory resources to users and lead to the overall inefficiency. This paper proposes a heterogeneous educational resource recommender system which is based on user preferences. It not only leads to a more efficient system that is targeted to user requirements but also tackles few of the major issues in most of the recommender systems such as the cold start problem. To tackle such issues, the system recommends latest trends to the user and learns from his behaviour towards these while the preferences are not set. One of the fundamentals of this evolving system relies on assigning a weightage to each recommended resource, calculated on user+s responses towards it. This is a vital ingredient in filtering out the non-relevant, non-informative and un-liked resources from being frequently recommended. The heterogeneous resource recommendation would leverage users in finding different types of relevant resources quickly to increase user productivity.