A Spatial Data Protection Algorithm Based On Differential Privacy

Nowadays the world-wide spread of Internet and the mobile devices have brought magnificent change to people's life.Due to the plenty of information hided by every mobile device user,it may produce billions of data exchange and data share events in the Internet every day.However,these events may cause endless privacy disclosure.Therefore,people's research on the direction of privacy protection and the exploration of the Internet are attached importance at the same time.The differential privacy model is a mathematically-based model that proposes a strict definition.Its privacy protection method is widely used in various fields of Internet protection.This paper mainly studies improved methods of spatial privacy protection based on differential privacy,eventually proposes two spatial data publishing algorithms based on differential privacy.The spatial data sets are divided into two types,and different protection algorithm mechanisms are given for different types of data sets.In the end,the availability of published data is improved while ensuring data privacy protection.This paper first proposes a selective perturbation algorithm based on differential privacy.After the algorithm uniformly adds noise to the spatial data set,a specific cluster area is selected by analyzing the data set.Therefore,the result of the first layer of uniform grid noise will be used by the data set,eliminating the privacy budget waste caused by the partitioning mechanism.At the same time,the classifying method based on differential privacy is used to make the data set avoid the uncontrollable force of equal noise accumulation for different distributions.In the case of privacy protection,the availability of data is improved.Secondly,after analyzing the advantages and disadvantages of the first algorithm,this paper proposes a Noise Added Selection algorithm.The algorithm divides the spatial data into uniform grids and then uses threshold filtering to perform true value selection and detection.The original data set is divided into blocks,and the data set that meets the threshold filtering conditions is adaptively protected by grid privacy.Integrate into coarse-grained data sets and protect privacy.The algorithm satisfies the diversity of spatial data distribution and ensures privacy protection for the same data.It not only provides advanced protection measures for densely distributed areas in spatial data,but also proposes corresponding protection measures for sparsely distributed areas.Furthermore,the availability of data can be ensured while ensuring privacy.We prove the availability of our methods by comparing experiment.