Abstract: An algorithm of double search mining based on transaction complementarity is presented，which is suitable for mining any long frequent item set.The algorithm uses top-down strategy and the way of transaction complementarity to generate frequent candidate item set in double search by binary logical operation，and uses transaction character to reduce the number of search transaction when computing support.Comparing with other algorithms，the experiment indicates that the algorithm is fast and efficient.

Key words: association rules, transaction complementarity, double search, binary, mining algorithm

摘要： 提出一种基于事务互补的双向挖掘算法，其适合挖掘任何长度的频繁项目集；该算法通过二进制的逻辑操作，用自顶向下策略和事务补集方式双向产生频繁候选项，并在计算支持数时用事务特征减少搜索事务数；通过与其他算法相比，实验表明它是快速而有效的。

关键词: 关联规则, 事务互补, 双向搜索, 二进制, 挖掘算法