The Research Of Privacy Protectionalgorithm In Data Publishing

With the rapid development of the information technology, more and more data are shared. In various kinds of network, people can easily collect many released information, even can collect a lot of information without privacy protection.This can easily violate individual privacy. People have paid more and more attention for these. How to make the released data carry individual privacy and can ensure the data availability so that people can dig out the knowledge has been the target of many experts.This thesis mainly based on the aggregation algorithm of k-anonymity model, specific work is as follows:Analysis the existing micro aggregation algorithm.Analysis MDAV algorithm which is fixed-length micro aggregation algorithm. In order to realize the k-anonymity algorithm which meet the l-diversity regular,this thesis proposed l-diversity-MDAV algorithm. The experimental results show that the algorithm can reduce the leakage risk.It can reduce the risk of getting information from the data table which is released by attackers.This thesis point out that the l-diversity-MDAV algorithm has a long operating time.The c-modes algorithm is introduced for this problem.This thesis combines the c-modes algorithm and MDAV algorithm and put forward the CLM algorithm. The algorithm clusters the data sets two times. Large-scale data set will be divided into several smaller data sets. This algorithm can effectively reduce the operating time, and can reduce the risk of leakage risk.This thesis analysis the MDAV algorithm which divise the equivalence group by the distance between tuples. The method is simple, but pay no attention to the intensive degree between recording tuples. Based on the MDAV algorithm, an improved algorithm is proposed. The improved algorithm uses Greedy algorithm and divises the equivalence group by the intensive degree between recording tuples.The improved algorithm can effectively reduce the information loss.The three algorithms this thesis presented improved the traditional k-anonymity algorithm from three aspects:The l-diversity-MDAV algorithm has matched the rules of l-diversity.CLM algorithm has improved the k-anonymity algotithm efficiently and the improved algorithm has significantly reduced the information loss.