Research On Location-Based Privacy Protection And Personalized Recommendation Methods

With the rapid development of mobile location technology and mobile social networks,people’s lifestyle has changed greatly.People begin to be more and more loyal to sharing their lives on the Internet,which directly or indirectly leads to the wide use of location-based social networks.In addition,through a large number of studies,it is found that analyzing a large amount of user location data and providing personalized location recommendation for users can not only find places of interest for users,but also improve the user experience of the recommendation system and win a good reputation for developers.Although accurate location recommendation can facilitate people’s life,the analysis and application of massive location information will inevitably lead to privacy leakage.Because personal location information can reflect a person’s life,work and rest,personal preferences,and even illness history.If these location information is obtained by illegal attackers,it will be very dangerous.Therefore,while personalized location recommendation,privacy protection of users’ location information is also an important part of the recommendation system.At present,in the location-based recommendation system,differential privacy has become the most effective strategy and method to protect users’ privacy data because it can resist strong background knowledge attacks.However,high-intensity privacy protection will also reduce data availability and low recommendation quality.Aiming at the above problems,this paper proposes a differential privacy protection method integrating distance weighting and social networks(DWSNP).Firstly,according to the distance between the user’s sign in position and the nearest sign in center,different levels of noise are injected into the user’s original sign in position;Then,according to the social relationship between the target user and his friends,add varying degrees of noise to the user’s friend location information,and finally achieve the purpose of protecting the user and friend location privacy.Through comparative experiments,it is verified that this method not only effectively protects the location privacy information of users and friends,but also improves the availability of data.In addition,the location data information based on privacy protection is more conducive to the user’s recommendation service.In recent years,the research on location recommendation methods has developed continuously,and the accuracy of recommendation system is getting higher and higher.However,most recommendation methods have some problems,such as the degree of personalized recommendation is not high enough,and a variety of influencing factors are not effectively integrated.Aiming at the above problems,this paper proposes a personalized location recommendation method based on multi factor fusion(MFFL).This method effectively integrates geographical location factors,location popularity factors,social relationship factors,time factors,weather factors and category preference factors,and fully considers the impact of various factors on the recommendation effect.Finally,through comparative experiments,it is proved that this method can effectively improve the accuracy,recall and F value of the recommendation system.