Research On Privacy Protection Technology For Wearable Device Data

With the rapid development and application of wearable devices,more and more user data that contains a large number of sensitive personal information is recorded and collected.It increases the concern about disclosure of privacy information,although providing users with rich personalized service.For the purpose of scientific research or data'sharing,a lot of data owners need to publish data to fully use the value of wearable devices.How to ensure the privacy in data publishing is one of the hot issues of current research.Firstly,this paper introduces the development of wearable devices and the basic knowledge of privacy protection technology,and then focuses on some key problems of privacy preserving data publishing process of wearable devices.This paper makes an improvement on Variable-Maximum Distance Average Vector algorithm,shorted for V-MDAV algorithm,by introducing two parameters:the weight W and the sensitivity S.The improved algorithm is WSV-MDAV and a new model for privacy-preserving data publishing is proposed based on the WSV-MDAV algorithm and differential privacy.The main work of this thesis is organized as follows:(1)This paper proposed a personalized distance measurement based on the weight W:W-PDM.When calculating the distance between records in traditional micro-aggregation algorithm,all quasi-identifiers attributes are assigned the same weight and the studies mainly focus on how to calculate the distance when there are multiple data types.This paper not only considers the individual needs of different users,but also considers the influence of the quasi-identifier attribute's distribution on the leakage of this attribute.The importance of different quasi-identifiers attributes can be better reflected by assigning different weights for them,and the distance between records,as well as the homogeneity of the micro-aggregation algorithm,can be more effectively measured.(2)This paper proposed the sensitivity S.This paper introduce the(?,k)method into the V-MDAV algorithm to address the attribute disclosure caused by homogeneity attack or similarity attack.And we set respective sensitivities S for different sensitive values,which is more flexible and provides better protection for the sensitive values.(3)This paper proposed a new model for privacy-preserving data publishing based on the WSV-MDAV algorithm:a data publishing model for wearable devices based on differential privacy.Firstly,the data is processed by the WSV-MDAV algorithm,and then the differential privacy is introduced into the data publishing process.On the one hand,the advantage of differential privacy is that it does not need to consider the attacker's background knowledge,and makes up for the defects of group-based model such as microaggregation.On the other hand,the microaggregated data reduces the sensitivity and the noise added to achieve differential privacy.Then the simulation and comparison of the proposed model and algorithm are carried out.The simulation results show that the proposed model and algorithm improve the privacy protection performance and reduce the information loss compared with the original V-MDAV algorithm,which can better realize the privacy-preserving data publishing of wearable devices.Finally,this paper summarizes the research results,points out the problems in the research process,and the future research directions are also discussed.