分布式数据隐私保护K-均值聚类算法  

Privacy-preserving K-means Clustering Algorithm over Distributed Data

作　　者：杨丹凤[1] 余青松[1] 郑冀之[1] 

机构地区：[1]华东师范大学信息学院计算中心,上海200062

出　　处：《计算机与数字工程》2008年第7期113-116,177,共5页Computer & Digital Engineering

摘　　要：如何获取准确的数据关系而不泄露合作方的任何私有数据是分布式数据挖掘隐私保护首要任务。将安全多方计算与数据挖掘技术相结合,提出应用于水平分布和垂直分布类型的数据的隐私保护k-均值聚类算法。实验表明算法能有效的保护数据的隐私,且对聚类结果没有影响。Privacy -preserving distributed data mining seeks to obtain accurate models without leaking the private data of the participated parties. Propose a privacy - preserving k - means clustering algorithm over horizontally and vertically partitioned data by integrating secure multi - party computation algorithm with data mining technology. Experiments show that it can efficiently preserves privacy in data items as well as guarantee valid clustering results.

关 键 词：分布式数据挖掘 隐私保护 K-均值聚类 安全多方计算 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]