Title: Improving the accuracy of item recommendations by combining collaborative and content-based recommendations: a hybrid approach
Authors: Desabandhu Parasuraman; Sathiyamoorthy Elumalai
Addresses: School of Information Technology and Engineering, Vellore Institute of Technology, Vellore, India ' School of Information Technology and Engineering, Vellore Institute of Technology, Vellore, India
Abstract: Recommender systems facilitate the users by providing the ample information of the items or the products they are interested. Users would not be aware of item details without the help of recommender systems due to the size of information available on the web. Collaborative filtering and content-based filtering are the two traditional filtering techniques of recommender systems. Both the filtering techniques have their advantages and certainly the disadvantages too. This can be solved by combining both the filtering techniques and improves the accuracy of recommendations. This leads to system as a hybrid recommender system. This paper presents a novel hybrid approach by combining a dynamic item-based collaborative filtering with the content-based filtering. Time variance and machine learning algorithms are applied on the filtering techniques to overcome the problems in recommendations. The approach is demonstrated using the MovieLens data sets to ensure the effectiveness of the proposed hybrid system.
Keywords: recommender systems; RS; collaborative filtering; CF; content based filtering; CBF; hybrid recommender system; HRS; machine learning; ML.
International Journal of Advanced Intelligence Paradigms, 2021 Vol.19 No.3/4, pp.262 - 270
Received: 19 Feb 2018
Accepted: 20 Apr 2018
Published online: 09 Jul 2021 *