Scenario of privacy violation within the recommendation databases

Recommendation systems are one of important systems. Currently, we can see recommendation systems to be embedded in several real-life systems such as Amazon, eBay, Facebook, Twitter, etc.,. Recommendation systems are designed to suggest or predict the eligible artifacts to target user, which recommendation systems always are based on the users' historical database, so called as the recommendation database. For recommending the artifacts, the recommendation system utilizes the existing information within the recommendation database in conjunction with the recommendation model to consider the eligible artifacts to target user. Indeed, the existing information within the recommendation database can be utilized by the recommendation agency people in some case such as mining the data to find a target group of users. With this reason, the individual users within the recommendation databases could be violated. Therefore, the objective of this work will be illustrating the scenarios of privacy violation that can occur in the recommendation database when the data holder allows the recommendation agency people to access the existing information within the recommendation database via the aggregate query functions.