面向互联网开放平台保护用户隐私的关联规则挖掘算法  

Study on privacy preserving association rules mining for Internet open platform

作　　者：王晗[1] 张玲[1] 

机构地区：[1]燕山大学经济管理学院,河北秦皇岛066004

出　　处：《计算机应用研究》2015年第5期1383-1386,1394,共5页Application Research of Computers

基　　金：河北省自然科学基金青年科学基金资助项目(G2011203195)

摘　　要：互联网开放平台提供的用户信息授权服务得到了广泛应用,但其在满足第三方网站的数据挖掘需求时往往将用户隐私信息交由多方存储,因而加重用户隐私滥用与泄露风险。针对这一问题,提出了一种开放平台与网站间的分布式关联规则挖掘算法,算法无须可信第三方参与,双方各自依据挖掘条件生成以频繁-1项集编号、用户身份标志符为行、列标记的布尔型矩阵,由开放平台进行矩阵扰动和整合,再由网站在整合矩阵上挖掘全局关联规则。实验证明该算法有效,且没有因通信代价而显著降低挖掘时间效率。In order to meet the demands of third-party websites in data mining process,the personal data authentication services provided by Internet open platform might add to security risk by storing privacy sensitive information on various websites.This paper proposed a distributed association rules mining algorithm for internet open platform and third-party website. The algorithm had no need for a trusted third-party,the two parties generate matrixes with the number of frequent-1 item set and user identification as markings of rows and columns based on mining conditions respectively,the matrixes of two parties were disrupted randomly and integrated by internet open platform. Website could generate global association rules based on the integrated matrix. Experiment shows that the algorithm is effective,and the mining efficiency is not significantly influenced by communication cost.

关 键 词：开放平台 隐私保护 分布式 关联规则挖掘 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP301.6[自动化与计算机技术—计算机科学与技术]