Research And Design Of Probabilistic Inference Based User Location Privacy Protection Algorithm

With the rapid development of the mobile communications and sensor equipment technology in the Internet Age,location based service become more and more popular,such as:cell phones built-in GPS and car navigation.We can get the accurate position information of some object by GPS even by WIFI.Data of the position information brings people great convenience.But at the same time,it also bring s harms for personal private information,because the location data contains both users' private location information and other sensitive information,such as users' individual habits,social status,health etc.In the widespread use of LBS,abusive location information can lead to loss of privacy.And we need a way to protect users' location privacy.In this paper,we analyze the users' location data from CMCC in Liaoning province.Then we use the Markov chain to model the users' mobile mode.Based on Probabilistic Inference and k-anonymity,we put forward some methods to protect users' location privacy.We inject the HMM into our location privacy protection algorithm.In the design process of this algorithm,we put k-anonymity into it,so that we can make sure that users' sensitive location could be protected well.While we are establishing this HMM mentioned above,we use the PSO algorithm to solve calculation procedure of obtaining the emission probability matrix.It takes few minutes to calculate it.However,we get a better performance than MaskIt.We also do some research on parallelized PSO,and at last we get a good performance in Hadoop environment which is based on MapReduce.As a result,the location privacy protection algorithm in the paper have got a good performance.Not only can we protect the privacy well,but also improve the quality of service while users use this method to protect their privacy.