基于FP-Tree的敏感性关联规则隐藏的研究  

Sensitive association rule hiding based on an FP-tree

作　　者：谢志强[1,2] 朱孟杰[1] 杨静[2] 

机构地区：[1]哈尔滨理工大学计算机科学与技术学院,黑龙江哈尔滨150080 [2]哈尔滨工程大学计算机科学与技术学院,黑龙江哈尔滨150001

出　　处：《哈尔滨工程大学学报》2009年第10期1134-1140,共7页Journal of Harbin Engineering University

基　　金：国家自然科学基金资助项目(60873037;60873019);黑龙江省自然科学基金资助项目(F200608);黑龙江省教育厅海外学人重点科研资助项目(1152hq08)

摘　　要：敏感性关联规则的隐藏是最大程度地保持原始数据集的其他特征,保证敏感规则不被挖掘出来.针对已有的基于对原始数据集中事务修改的方法产生大量I/O操作的问题,提出了基于频繁模式树(FP-tree)的敏感性关联规则隐藏的方法.该方法首先利用FP-tree存储了与事务数据库相关的全部信息,减少了产生和测试候选集耗费的大量时间;再利用改进的频繁模式树(IFP-tree)是单向的,快速挖掘出最大频繁项目集,确定敏感性关联规则;然后删除敏感关联规则对应的频繁项目集,更新IFP-tree项目集节点和相应的项目头表的计数,对更新的IFP-tree反向挖掘生成新的不包含敏感关联规则的事务数据库.实例和理论分析表明,该方法是正确和高效的.Hiding, using the sensitivity association rule, is done to maintain characteristics of primitive data sets, with the condition that its use should be transparent. The existing method that was based on changes made to original transaction data generated massive I/O operations. So an effective method for hiding which makes use of sensitivity association rule based on FP - tree was proposed. This method using the FP - tree to store all of the information related to the transaction database greatly reduced time for the generating and testing of a candidate set. As the improved FP - tree （ IFP - tree） is one - way, it can be used to rapidly mine the maximum of frequent item sets and determine their sensitivity association rules. Then frequent item sets which support the sensitivity association rules are deleted ; IFP - tree item sets＂ nodes and the count of corresponding item head tables are updated. Finally, a new transactional database that excludes sensitivity association rules is generated by reversely mining the updated IFP- tree. Examples and theoretical analysis show that this method is accurate and efficient.

关 键 词：数据挖掘 敏感关联规则 关联规则隐藏 频繁模式树 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]