Font Size: a A A

Location Privacy Protection Technique Based On Visit Probability

Posted on:2018-12-15Degree:MasterType:Thesis
Country:ChinaCandidate:D P ZhaoFull Text:PDF
GTID:2348330512481301Subject:Software engineering
Abstract/Summary:PDF Full Text Request
In recent years,the researchers suggest a lot of new methods for location privacy protection.The existing technology seldom utilizes side-information,such as visit proba-bility,map,location semantic,which are likely to be exploited by malicious attackers for users' location privacy.Even if some researchers take side-information into account,they model the side-information universally for all users rather than personally for every user.In this paper,we propose two methods to protect the users' privacy in location-based ser-vice(LBS).We select the dummy locations to achieve k-anonymity according to global visit probability and personal visit probability for users' queries.Main contributions of this paper are as follows:· Location Privacy Protection Method Based on Global Visit Probability Global visit probability based on location privacy protection first consider visit feature from majority of users,that is,global visit probability for different places in large amounts of users' visit data.According to users' anonymity requirement k,we re-trieve dummy zones meeting users' demand,calculate location information entropy for each area,select m areas with greatest information entropy and select one area for users' dummy location randomly.· Location Privacy Protection Method Based on Personal Visit Probability We propose a personalized visit probability based on location privacy protection,namely EPLA.For any user,we calculate a set of distance between his past visited locations.We estimate the personalized distribution for user's location distance using Kernel Density Estimation(KDE),thus calculate the personalized probability for each loca-tion anchor.Then we perform user's personalized location anonymity by selecting proper candidate anchor using ASS.·Optimization Algorithm of EPLA To optimize EPLA,we use fast Gaussian rule and 3?-principle Gaussian distribution by KDE computing feature and propose an efficient location privacy method APLA,which optimizes EPVP algorithm in EPLA.APLA greatly reduces the computational complexity.
Keywords/Search Tags:Location Privacy, Location-based service(LBS), k-anonymity, Visit probability
PDF Full Text Request
Related items