Toward Privacy Protection for Location based Recommender Systems: A Survey of the state-of-the-art

Abstract

Recommender systems help users to find and evaluate their point of interest (POI) where the number of choices can be overwhelming. These systems incorporate data mining techniques and make their recommendations using knowledge learned from past experience of user's actions and attributes. Recommender systems are categorized based on services, information and recommendations for the users. Location based Recommender System (LbRS) is one of these systems where the user requires the recommendations for his/her point of interest (POI). In order to get the desired recommendations, LbRS imposes to reveal personal information along with the current location. However, revealing the user's profile allows bringing out many aspects of one's personal life that raise many privacy issues and decrease frequent usage of recommender systems. In this paper, we identify the most concerning privacy metrics that are required to be protected in LbRSs. In addition, we demonstrate the situations when these metrics are required to be protected and not. Furthermore, belonging to these privacy protection metrics, we discover the different privacy attacks that can be encountered during query processing for getting desired recommendations. Moreover, a comprehensive study was conducted and presented different privacy protection approaches and their concerns to protect these privacy metrics.