Reasearch On Privacy Protection Based On K-anonymit Technology

With the development of the electronic information technology, data mining technology andmobile positioning technology, obtaining large amounts of information becomes possible. However,while people enjoy the service of freely searching information, people are also faced with the risk ofprivacy disclosure. K-anonymity, is a widely used technology in data and location privacyprotection. It can meet the requirements of privacy protection in data release and location basedservice. In data release, the K-anonymous technology protects data privacy at the expense of theoriginal data quality. And in location based service, K-anonymous technology achieves the purposeof privacy protection through bluring some user’s location. A K-anonymity algorithm with highperformance can make users avoid suffering unnecessary losses caused by privacy disclosure indata publishing and location based service. Therefore, it has an important theoretical value andpractical significance to research into K-anonymity technology.This paper researches the K-anonymity technology, used in privacy protection for both datarelease and location based service. After thoroughly analyzing the existing K-anonymity technologyand the possible attacks on it, we propose two new K-anonymity algorithms. The main work andachievements are as follows:1. The concept, features, possible attacks, research status and research hotspot of theK-anonymity technology are introduced in detail. And its common implementation methods andanonymity metrics are studied.2. Based on the study of MDAV algorithm, a new data privacy protection algorithm fordynamic data release is proposed. This algorithm can be used in the dynamic data release scenariowhere the traditional algorithm cannot. The algorithm inherits the distance oriented ideology ofMDAV algorithm, and gives a method to calculate the distance between the non numericalattributes. Using the calculating distance, the degree of similarity between attributes can bemeasured. In the process of calculating distance, not only the hierarchical relationships betweenattributes are considered, but also the statistical characteristics of attribute frequency in the wholedata set. Thus a good measurement of the relationship between non numerical attributes among thewhole data set can be made.3. In order to overcome the disadvantages of imprecise query results and high communicationoverhead that exist in traditional anonymous scheme, a new kind of location anonymity scheme isproposed. This scheme divides single cloaking region into several sub cloaking regions, andreplaces the users’ real position with the central location of sub cloaking region which the users belong to, then initiates a query from LBS server. Compared with traditional scheme, this one canensure that users achieve accurate results relatively and minimize the communication traffic.4. Using entropy theory, we make an assessment of the location anonymity algorithm. Theentropy is calculated on the condition that the probability of the leakage of users’ privacy isregarded as the event probability. Using the calculated entropy to measure the degree of privacyprotection, the greater the entropy is, the higher the degree of privacy protection will be.