关联规则挖掘中的隐私保护研究

Research on Privacy Preserving in Association Rules Mining

数据挖掘中的关联规则反映一个事件和其他事件之间依赖或相互关联的知识。随着大量数据不停地收集和存储积累,人们希望从中发现感兴趣的数据关联关系,从而帮助他们进行决策。随着信息技术的发展,数据挖掘在一些深层次的应用中发挥了积极的作用。但与此同时,也带来隐私保护方面的问题。隐私保护是当前数据挖掘领域中一个十分重要的研究问题,其目标是要在不精确访问真实原始数据的条件下,得到准确的模型和分析结果。为了提高对隐私数据的保护程度和挖掘结果的准确性,提出一种有效的隐私保护关联规则挖掘方法。针对关联规则挖掘中需预先给出最小支持度和最小置信度这一条件,提出了一种简单的事务数据库中事务的处理方法,即隐藏那些包含敏感项目的关联规则的方法,以对相关事务作处理,达到隐藏包含敏感项目的关联规则的目的。理论分析和实验结果均表明,基于事务处理的隐私保护关联规则挖掘方法具有很好的隐私性、简单性和适用性。

Association rules mining in data mining reflects relations between events. With the large amounts of data collection and storage continuously accumulated, people want to find data associations which they are interesting in, and to assist them in decision-making. With the developments of information technology, data mining plays an active role in applications. But at the same time, it has brought some problems of privacy preserving. Privacy preserving is currently a very important issue in the field of data mining. The object is to get veracious model and analyze the results with imprecise data acess. In order to raise the level of protection of data privacy and the accuracy of mining results, propose an effective privacy preserving method. The minimum support and confidence should be given in associations mining,against this,a simple transactions handling method has been given. Can hide the associations which contain sensitive items by the way of dealing with transactions. Theoretical analysis and experimental results show that, this method based on tansaction processing has got good privacy, simplicity and applicability.