Long-term User Engagement in Recommender Systems: A Review

Abstract

The use of recommender systems is considered to be highly significant in our day-to-day life. The feed streaming approach has been used to a considerable degree inside the recommender system because of its effectiveness. Through the use of the feed streaming setting, users are given the opportunity to engage in an interactive manner of recommendation inside never-ending feeds. The concept of user stickiness, which goes beyond the conventional concept of instant analysis and is often evaluated by long-term user involvement, should get a higher amount of attention from a powerful recommender system. Specifically, this is due to the fact that user stickiness is not within the purview of conventional measures. A primary objective of recommender systems (RS) is to improve the level of user participation and communication that occurs inside the platform. In spite of this, there is a lack of comprehension of the extent of this link and the ways in which RS might potentially improve continuous user engagement with the platform to a certain degree. The present study endeavours to analyse the role that RS plays in turning users' short-term engagement with the RS into engagement with the platform that is maintained over time. This is done in order to solve the knowledge gap that has been identified. In order to investigate these concerns, we first construct a conceptual framework by doing a literature research on the relevant literature in the fields of user engagement and recommendation systems. This is done in order to get the necessary information. Within the scope of this research, we talk about metrics for the evaluation as well as open challenges that are presently being encountered in this aspect of the field.