A Survey of Serendipity-Based Recommendation Systems

Many recommendation systems have been researched and published in the recent decade by many researchers and scholars. Many recommendation systems have been commercially used by those commercial companies such as digital music providers like iTunes, e-commerce website like Amazon, YouTube, a social networking provider like Facebook, and etc. The reasons in the rapid growth of the number of recommendation systems are due to the advancement of the Internet and technologies such as mobile devices and database. Recommending something that users are familiar or known is no longer enough to satisfy the users nowadays. Many researchers are now looking into ways to recommendation something that are out of the box, and that will give a pleasant surprise to the users. In this paper, a survey on the development of serendipity-based recommendation systems in recent time will be conducted. The survey will look into the mechanisms and approaches used in recommendation systems. Then various user model, data model, and algorithms will be discussed. Besides that, a number of serendipity-based recommendation systems, together with its evaluation approaches, will be surveyed. Lastly, some challenges in implementing serendipity-based recommendation systems will be discussed. Keywords- serendipity, recommendation system, collaborative filter, user model, data model, content-based model, context-based model.