Sensitivity-based (p, α, k) - Anonymity Privacy Protection Algorithm

Abstract

Medical data itself has extremely high research value, but how to protect its privacy and security in the process of sharing medical data has attracted widespread attention from researchers. Aiming at the problems of homogeneity attack, background knowledge attack and high-sensitivity similarity attack in data sharing of k -anonymity privacy protection algorithm, a sensitivity-based (p, α, k) -anonymity privacy protection algorithm is proposed. The concept of semantic similarity tree is introduced, which can resist background knowledge attacks. The improved clustering method of equivalence classes can solve homogeneity attacks and high-sensitivity similarity attacks. Thus, the security of medical data sharing can be realized. Experiments show that (p, α, k) - anonymity privacy protection algorithm has the best performance when α is equal to 0.5. In addition, compared with k -anonymity privacy protection algorithm, although (p, α, k) - anonymity privacy protection algorithm has higher execution time and information loss, it effectively solves the problems of k - anonymity algorithm and improves the security of medical data sharing.