Research And Implementation Of K-anonymity Privacy Protection Algorithm Based On Local Generalization

In the era of information,while data sharing provides convenience for people,it causes serious personal privacy disclosure.Therefore,how to protect personal sensitive information in data is becoming increasingly important.In the field of information security,k-anonymity privacy protection technology in data release realizes the anonymization of data by breaking the one-to-one correspondence between the data owner and the data record in the original dataset,and protects the sensitive information in data,which effectively reduces the risk of personal privacy information disclosure in the released data.Therefore,this paper will research k-anonymity privacy protection algorithm based on local generalization in data publishing.Firstly,this paper introduces the background and research status of privacy protection technology,and analyzes the mainly anonymous models in privacy preservation of data release.Then it introduces the definition and some theoretical foundations of k-anonymity.The research on implementation methods and algorithms of k-anonymity is investigated and summarized in this paper.Secondly,this paper designs a privacy protection scheme GH-SPK,and presents RC-OTF algorithm aiming at the problem of unnecessary information loss caused by attribute generalization hierarchy that constructed by OTF algorithm in anonymization.The algorithm creates attribute generalization hierarchy by classifying numerical attributes and considers node frequency in hierarchy when a new node is created,which reduces the information loss caused by generalization hierarchy in anonymization.Then the correctness and effectiveness of RC-OTF algorithm are proved by comparative experiments of three classical generalization hierarchy algorithm.Finally,in order to solve the disclosure of sensitive information caused by homogeneous attack and background knowledge attack in the anonymous dataset realized by Bottom Up local generalization algorithm,this paper proposes an improved k-anonymous local generalization algorithm SPBU.The algorithm constrains the frequency of sensitive values in the equivalence class used by the sensitive attribute generalization tree,and meets the requirement of personalized privacy protection of users by adding guard node for sensitive values.At last,the recognition rate of sensitive value and information loss are used as metrics and attribute generalization hierarchies are created by RC-OTF algorithm to prove that SPBU algorithm can reduce the risk of privacy information disclosure by comparative experiments.