Research On Privacy Protection Method Of Multi Sensitive Attributes Data Set Oriented To Mashup

As a new web-based data integration application, Mashup is springing up on the Internet.It provides strong support for exchanging and sharing data. The connection operation among these data sources often leads to serious privacy disclosure,also reveals the sensitive information to other data providers.And the mashup data would suffer from the problem known as the curse of high dimensionality, resulting in useless data for further data analysis.Privacy protection in data mashup is important and challenging.PHDMashup algorithm is proposed to deal with the problems in data mashup.lt uses LKC-privacy model and combines top-down specialization approach to achieve privacy protection.However,many data providers are involved in the process of data mashup,and the number of attributes to be anonymized is huge. The PHDMashup algorithm requires to specialize all nodes in the generalization trees of all the attributes. It not only causes the waste of time and space,but also brings a heavy computational load.In this paper,NPHDMahup algorithm based on PHDMahup is proposed .It improves the efficiency of the algorithm by reducing the specialized nodes. Moreover,considering that there is a lot of time waste because of the communication among data providers due to the above two algorithms,an improved data mashup algorithm with high efficiency for privacy protection named SPHDMashup is proposed.By introducing a server as the middleware,not only the efficiency of the algorithm is enhanced, but also the workload of data providers is greatly reduced.Moreover, a way to solve the heterogeneous problem of the data is also proposed.Finally,for the proposed algorithm, experiments are carried out and verify the superiority through the analysis and comparison with the original algorithm.The shortcomings and improvement direction in the future of the algorithm are also discussed.