Research On Location Information Protection Based On Geographic Indistinguishability

Differential privacy has been widely recognized by academia since it was proposed in 2006,and has been gradually applied to social networks,histograms,data mining,location privacy protection,and other scenarios.Differential privacy mainly provides privacy protection for onedimensional data.When it is applied to the field of location privacy protection,the key and difficult point is how to realize the transformation of application scenarios from one-dimensional space to two-dimensional space under the condition of satisfying differential privacy.The emergence of geoindistinguishability in 2013 filled in the gap of differential privacy in the application of geometric space.However,the existing privacy protection mechanism of geo-indistinguishable location has some problems and deficiencies in query accuracy,noise,and privacy budget allocation.Therefore,how to effectively combine specific privacy protection scenarios and design location privacy protection mechanism to meet the geo-indistinguishability is the key to balance the accuracy and effectiveness of the mechanism.This paper mainly focuses on the following problems: the small-scale retrieval error of geoindistinguishability algorithm is large,the privacy protection of users' frequent check-in locations,the quality loss of privacy and the privacy measurement.The research content mainly involves the location privacy protection mechanism that meets the geo-indistinguishability,the strengthened frequent location protection method and Application based on geo-indistinguishability.The specific work is as follows:(1)Privacy protection mechanism of geographic indistinguishable location basedon the user's region of interest.In order to solve the problem that the existing geo-indistinguishability algorithm has a large error in the results of small-scale queries,a location privacy protection model that satisfies the geo-indistinguishability is constructed by taking the region of interest as the privacy constraint.Secondly,the geo-indistinguishability algorithm is improved by using the model.Finally,experiments based on real data show that the privacy protection algorithm based on user's region of interest can improve the accuracy of small-scale query results without reducing the level of privacy protection.(2)This paper proposes an algorithm based on geo-indistinguishability to enhance frequent location privacy protection.The algorithm classifies the location points into clusters with different frequencies according to the access rate of the location.In order to protect the location set visited by users and hide the sensitive information of the high access area,the algorithm introduces the privacy budget allocation factor to guarantee the privacy of different clusters in different degrees.The goal is to blur the high access area and make it difficult to distinguish it from the low-frequency area.(3)This paper proposes a method to measure the amount of privacy disclosure.From the attacker's point of view,we consider the user's historical trajectory,the query function of the disturbance mechanism,and the privacy measurement method of the mechanism.On this basis,we introduce information entropy to measure the amount of information that may be helpful for the attacker to infer.