A goal-based hybrid filtering for low-rated users recommendation issue using neighborhood personalized profile similarities in E-learning recommendation systems

Abstract

The e-learning recommender systems are based on the users past history, ratings, likes or dislikes. The low-rated (less rated history profile) users may cause of zero or non-relevant recommendations issue in these days, which lose the users interest. This research proposed the goal-based hybrid filtering approach that used to perform the personalized similarities between users personalized profile preferences collaboratively. The aim of this research study is to improve the low-rated user’s recommendations by tackling the collaborative filtering and k-neighborhood personalized profile preferences similarities in e-Learning recommendation scenarios. The experiments has been tackled with famous ‘Movielens’ dataset while the experimental results has been performed with the help of (average mean precision Pr: 79.90%) and (average mean recall Re: 83.50%) respectively. A conducted result demonstrates the effectiveness of proposed goal-based hybrid filtering in the improvement of low-rated users profile recommendations in e-learning recommendation systems.