垂直分布下的隐私保护关联规则挖掘-计算机技术专业论文.docxVIP

摘 要 近年来，数据挖掘作为一种在实际生活中广泛应用的大数据智能处理技术，给人 们带来便利和经济效益的同时，不可避免地给人们的隐私带来不安全因素，造成隐私 泄露的问题。隐私保护数据挖掘作为一个新兴的研究方向，被越来越多的学者所关注； 同时隐私保护数据挖掘中的隐私保护关联规则的研究，也逐渐成为数据挖掘领域的热 点研究问题之一。 本文围绕垂直分布式环境下的隐私保护关联规则挖掘进行研究，主要工作和内容 如下： (1) 围绕隐私保护关联规则的相关内容进行阐述，分析已有隐私保护关联规则挖 掘算法的优缺点，同时总结支持垂直分布关联规则挖掘的隐私保护技术以及影响关联 规则挖掘效率的几个重要因素，这些因素同样也影响着隐私保护关联规则的挖掘效 率。 (2) 针对以往垂直分布下的隐私保护关联规则挖掘算法所存在的不安全性和不 高效性，提出隐私保护最大频繁项目集挖掘算法，同时设计一个新的点积协议，并将 其应用到算法中。利用 JAVA 语言实现算法，验证了算法的高效性。 (3) 针对隐私保护最大频繁项目集挖掘丢失了某些频繁项目集的支持度以及两 个参与方对算法的限制性，结合隐私保护的多方安全集合交集基数协议，提出了一个 多方参与方的、垂直分布下的隐私保护频繁闭合项目集挖掘算法。算法开创性提出了 垂直分布式环境下由局部频繁项目集产生全局频繁闭合项目集的项目集生成方式，降 低了计算代价，并通过实验验证了算法的可行性。 关键词：垂直分布；多方安全计算；最大频繁项目集挖掘；频繁闭合项目集挖掘； 隐私保护关联规则挖掘 - I - Abstract In recent years, data mining brought the leakage of the privacy while it brought benefits to people. To solve this problem, privacy preserving data mining was came up and quickly became the hotspot in data mining research. The association rules mining is an important task of data mining. So more and more scholar focused on the privacy perserving association rules mining. This dissertation extensively studied algorithms for privacy preserving association rules over vertically partitioned data. The main content, contribution and innovation in the dissertation are described below. We found out the advantages and disadvantages of the existing privacy preserving association rules mining algorithm by studying the contents of the privacy preserving association rules mining. And we also summarized the privacy protection techniques which can support the association rules mining over vertically partitioned data and the key influencers about the association rules mining. These factors were the key factors in the privacy preserving association rules mining. To overcome the shortcoming of privacy preserving association rules mining over vertically partitioned data, two new privacy preserving association rules mining algorithms were presented in this paper. The first algorithm was about the privacy preserving maximal frequen