Improving top-K contents recommendation performance by considering Bandwagon effect: Using Hadoop-Spark framework

Abstract

The study on the existing Collaborative filtering recommendation system is mainly aimed at improving the accuracy of prediction. However, in terms of actual recommendation service, it is more important that the Top-K recommendation list, which is effectively recommended to the user, is an item that the user actually likes, rather than improving the recommendation accuracy of all items. In this paper, we have developed a recommendation system that considers the psychological concept of Bandwagon Effect in order to improve the recommendation accuracy of the Top-K contents. For Big data distribution and storage, we used Hadoop and for the fast Big Data processing offering speed, we used Spark, an in-memory data processing framework for high-speed operations. As a result, the proposed model is superior to the existing model in terms of accuracy of recommendation for Top-K contents.