基于K匿名的隐私保护算法研究-计算机系统结构专业论文.docxVIP

ABSTRACT 丁oday，with the rapid development of the information age，more and more data are provided for people tO share and use，at the 盹me time more and more attention has been paid for the privacy protection of the released data. So when r，佣lizing 侃侃， we should both protect the privacy of the data and ensure data integrity as much as ‘ possible ，then the 巾taωn be used for people. A method called K-anonymity was propos创.K-anonymity is a technique 伽t prevents joining attacks by generalization or suppressing portions of the source data 50 that no individual can be uniquely distinguished from a group of size k. K-anonymity is simple and is 佣sily to be implemented; it also ωn reduce the information loss at the maximum extent at 由e 制me time protecting personal privacy. Ofa11 the Kanonymity algorithms，由ere is an efficient algori也m called optimal la饥ice anonymization(referr 创 to OLA) algorithm. It uses a structure called lattice. We get the optimal result t加Ough traversing the nodes of the lattice. But the traver时 。rd町 of lattice nodes of the OLA algorithm ωnnot reduce the lattice nodes at the most 也at need to be computed， which increase the run time Of algorithm， and because of 也e existence ofthe i50lated data in the ωurce data ，we 铀的 get a good .result on the information loss through processing the source data only once using the global nonymity algorithm. In 也is paper we improve the OLA algorithm on these two a叩ec邸， and these improv创nents have better results. Of the next 出ree points，the :first two are directed to reducing the run time of the algorithm and the last one is directed ωreducing the information loss ofthe processed data: 、?To the traversal sequence of the lattice nodes，we proposed a me由od which the traver础1 sequence is due to the product of d嗨ree of the node. Experiments show 巾at 让 canreduce 伽e lattice nodes which should be computed，which de町ease the run time of the algorithm. 、 ? 丁。由e number of the lattice nodes to be computed， we u