Personalized Location Recommendations With Privacy Protection

With the development of Internet technology,location-based social networks(LBSN)is also widely used,users are willing to share their lives and personal preferences on the Internet.Now,the popular Weibo and WeChat are enough to prove the hot of mobile social networks.Research and analysis of huge user's location information data,it provides users with personalized location recommendation services,and it helps users to discover and visit location areas that they have not touched,which is a place will interest them.Not only to avoid users wasting time thinking about where to go,it also allows users to know where to go to make themselves happy.Due to the extensive use of location information data,users are paying more and more attention to the privacy issues of personal location data information.Location information data can highlight the user's activity patterns,the preferences of visiting places,and even the rules of life,which can be easily obtained or inferred by an attacker.Protecting the user's location privacy data is an important part of the location recommendation.Whether it is the user's original check-in information or the calculation involves user's privacy in the recommendation process,they are all the object to be protected.This article has conducted in-depth research and exploration based on personalized location recommendations with privacy protection,including the following points.1.In order to cover up the user's original check-in data information list and prevent the attacker from using the user's other relevant personal information to infer the privacy information,this paper will combine the two methods of differential privacy and random perturbation.Considering multiple levels of random perturbation methods,different ranges of perturbations are applied to different user's check-in frequencies to achieve privacy protection personalization,and also to improve the confidentiality of data.Then add noise generated by Laplace distribution to friend relationship calculation.In order to satisfy the differential privacy,it is used to prevent an attacker from inferring the user's friend information.2.The previous research data analysis shows that the data scrambling method to achieve the privacy protection effect will reduce the recommendation accuracy of the recommendation system.Based on the different characteristics of the user information,this paper proposes a personalized location recommendation.The information is divided into two types: privacy protection information and non-privacy protection information,so that some user's information with risk of privacy leakage will be effectively protected,and information without risk is mainly used to improve the accuracy of recommendation.The results of experiments show that the method can protect the user's private information,and also provide users with accurate recommendations.