Personalized content recommendations on smart TV: Challenges, opportunities, and future research directions

Abstract Web recommender systems play a significant role in different domains, such as movies, books, music, etc., and contributes to not only user satisfaction but also to e-business and e-commerce. It utilizes the user’s profile information, preferences, and activities for recommendations of different objects. However, the distinct nature of watching smart TV greatly affects the accuracy and efficiency of recommender systems due to the reasons that it is a lean-back and multi-user device. Hence, the predictions and calculations of a user’s profile, preferences, and activities cannot be accurately utilized by the recommender systems for recommendations to the exact viewer(s) watching smart TV. This paper presents a critical review of existing recommender systems in the perspectives of smart TV watching scenarios. It highlights the issues and challenges and presents some research opportunities to deal with it. It further presents a subjective study for validating the highlighted factors that affect the recommendation results specifically on a smart TV. Results show that watching activities on a smart TV is significantly different from other devices, such as smartphones and computers. It further shows that smart TV is a non-personalized device and normally enjoyed in groups. Hence, personalized recommendations on smart TV need further investigation. The paper concludes that the existing recommender systems need further investigation to cope with issues of recommendations on a non-personalized device i.e., smart TV. Improving the recommender system for smart TVs may contribute not only to the viewer’s satisfaction but also to the conversion rate.