Research On Privacy Protection Algorithm Based On (?, K)-Anonymity

In recent years,the rapid development of the Internet,the network provides a platform for information exchange and data sharing,data holders to publish data on the Internet,and researchers will select the data in a large number of data.With the attention of the user's privacy information,privacy protection has become an important issue in the process of data publishing.On the one hand,when the data holder data released after treatment,need a certain degree of privacy data were released;on the other hand,has released data,while the data analysis,we should maximize the availability of data.Now there are many kinds of privacy protection model,for example: k-anonymous model,(?,k)(?)-anonymous model,-diversity privacy protection model,at the same time to protect the privacy of users,but also to a certain extent in the data usability.k-anonymous model data released by the low precision,vulnerable,and relative to the k-anonymity algorithm,(?,k)-algorithm of data anonymous protection degree higher,were studied in this paper based on(?,k)-anonymous privacy protection algorithm,the main work is as follows:(1)Through the research on the present situation and research significance of the domestic and international privacy protection,the paper analyzes several classical privacy protection algorithms in the field of privacy protection.(2)An(?,k)-anonymous algorithm based on differential privacy is proposed.In order to meet the requirements of(?,k)-anonymous data tables,the database query table is established,which is composed of the records corresponding to the query conditions.In order to protect the privacy of users,reduce the risk of leakage of user privacy,while improving the quality of data.(3)The study of trajectory privacy data based on Mean Shift algorithm.In the process of user movement,there will also be a privacy leak,this paper will use the Mean Shift algorithm in the user's path privacy protection.First of all,the user's moving trajectory is divided into different subsets by preprocessing the sampled data.Secondly,we use the Mean Shift vector to find the mean value of each trajectory in the sampling time.Finally,the Mean Shift algorithm is used to cluster the trajectories,and the trajectory is(?,k)-anonymously released.(4)Through the experiment,the two improved algorithms and the existing algorithms in the time efficiency and information loss to make a comparison,and gives the experimental analysis.Based on(?,k)-the anonymous privacy protection algorithm,the algorithm is analyzed and improved in detail,proved by the experimental results and the performance analysis,the improved algorithm has greatly improved in terms of privacy protection and data availability.