| With the rapid development of computer network technology and the rapid increase of storage equipment,the collection,storage and filtering technology of massive data are becoming more and more sophisticated.In order to achieve mutual benefit and win-win,people will collect the data published on a public platform for relevant personnel to carry out data mining and data research.But these data contain a lot of information related to personal privacy,if there are illegal access to these private information,will endanger the interests of individuals or groups,and even threaten the personal safety of users.Therefore,while enjoying the benefits of information globalization,we should also consider how to effectively protect the privacy and security of users.The most common approach to privacy protection is to remove information that identifies an individual,such as an identity card number,and then publish that data as a table.However,if an attacker obtains the user's personal information table with identifier from other ways,the private information of the user will be obtained by link attack.In addition,background-knowledge-based analysis may also result in privacy leakage.Generalization can resist link attack very well,but after generalization processing,data availability will be greatly reduced.How to ensure the real availability of published data without compromising privacy has long been a concern of researchers.The early K-anonymity technology uses generalization to prevent link attack and L-diversity anonymity technology to help prevent homogeneous attack,but generalization technology will cause serious loss of information.In view of the availability of data,anatomy separates the standard attribute from the sensitive attribute to replace the generalization technique,which guarantees the real validity of the quasi-representation attribute and makes the published data more valuable for research.But this method is very weak to the sensitive attribute privacy protection measure,easily causes privacy leakage.In addition,based on the needs of users,some standard attributes need to be protected accordingly,and different types of sensitive attributes are valued differently by users.Firstly,this paper analyzes and compares the shortcomings of K-anonymity and L-diversity anonymity,investigates and studies their defects,and makes some adjustments to these two algorithms based on generalized hierarchical tree: 1.According to the function degree of the standard identification attribute to the sensitive attribute,the different standard identification attribute is generalized in different degree,which makes it more flexible to protect the user's data privacy.2.According to the difference of privacy importance of sensitive attribute,the sensitive attribute of equivalent group is defined,and the size of equivalent group and the classification of sensitive attribute are limited by threshold.Secondly,after studying the core idea of anatomy method and comparing the advantages and disadvantages of anatomy method,a new mapping-based privacy protection algorithm is proposed.This method solves the problem of information loss caused by data generalization,it also guarantees that the privacy of sensitive attributes can not be easily divulged,and gives the concrete algorithm steps.Through experiments,the advantages and disadvantages of the new algorithm in function and performance are analyzed.Finally,a new algorithm is proposed based on the above two points:(F,?)anonymity algorithm based on multi-dimension sensitive attribute privacy protection,and the algorithm design and main structure are given,it can not only protect the privacy of individuals but also improve the accuracy of information,and it can protect some quasi identity attributes and sensitive attributes flexibly according to user needs.At the same time,through experiments,the(F,?)anonymity algorithm based on multi-dimension sensitive attribute privacy protection is compared with other algorithms to prove its value. |
PDF Full Text Request