Abstract: This paper introduces the idea of T-testing into privacy preserving data mining algorithms, proposes privacy preserving association rule mining algorithm based on influence measure. Considering influence measure as association rules generated as a criterion is to reduce the redundant rules and irrelevant rules so as to improve the efficiency of mining. Sensitive rules can be hided by adjusting the transaction association rules between the sensitive rule hiding sensitive items to achieve. Experimental results shows that, the algorithm makes the rules for side effects such as loss rate and the rate of decrease to as low as 6%.

Key words: privacy preserving, association rule, influence measure, data mining, sensitive rule

摘要： 将T检验思想引入隐私保护数据挖掘算法，提出基于影响度的隐私保护关联规则挖掘算法。将影响度作为关联规则生成准则，以减少冗余规则和不相关规则，提高挖掘效率；通过调整事务间敏感关联规则的项目，实现敏感规则隐藏。实验结果表明，该算法能使规则损失率和增加率降低到6%以下。

关键词: 隐私保护, 关联规则, 影响度, 数据挖掘, 敏感规则

CLC Number: 

• TP182