Review of User Interface-Facilitated Serendipity in Recommender Systems

Abstract

Serendipity has become a desirable quality in designing recommender systems and user interfaces, hence offering a new measurement for system quality. At the same time, recognizing serendipitous experiences and determining their value is difficult due to their subjective nature. This review builds on 10 studies of user interfaces facilitating serendipity studies and attempts to understand the patterns that guide relevant user interface designs in recent recommender systems. This study furthers our understanding of turning the elusive experience of serendipity into more actionable user interface designs and patterns. The key findings are as follows. First, user controls and visualizations have facilitated serendipity, but studies of recommender systems have not gained considerable attention. Second, frameworks instrumental for user-interface-facilitated serendipity have not gained the researcher's worthy attention. Third, developing countries need to explore serendipity-facilitating recommender systems with more diverse users and more prominent test cases.