摘要
在数据发布中的隐私保护研究中,实现有损连接的方法主要有基于匿名模型方法和基于贪心策略的(α,k)匿名聚类方法。针对基于匿名模型方法存在的效率低以及基于贪心策略的(α,k)匿名聚类方法得到数据有效性差等不足,提出基于相似度的有损连接方法,该方法根据发布数据之间的相似性聚类得到有损连接的结果,解决了目前有损连接方法所存在的效率以及准确率问题。实验结果表明,该方法能够有效实现发布数据的隐私保护。
In privacy preservation research with regard to data publishing,the methods to realise lossy join are mainly the anonymous model-based approach and the greedy strategy-based（α,k） anonymous clustering approach.The former has the problem of low efficiency and the latter has the deficiency of poor effectiveness in data derived.In light of the drawbacks of these two methods,we present a similarity-based lossy join method,the method obtains lossy join results according to the similarity clustering between the published data,this resolves the problems of current lossy join methods in efficiency and precision.It is proved by the experiments that this new approach can effectively realise privacy preservation in data publishing.
出处
《计算机应用与软件》
CSCD
北大核心
2012年第11期123-125,242,共4页
Computer Applications and Software
基金
教育部人文社会科学项目(11YJC870011)
北京市教委科技计划面上项目(KM201211232014)
北京信息科技大学网络文化与数字传播北京市重点实验室开放课题(5026035409)
