摘要： 针对数据库不断更新变化及现实生活中大多只对近期数据感兴趣的特点，该文提出了一种基于滑动窗口过滤器的关联规则增量式挖掘算法（SWFAI算法）。该算法通过分组及时舍弃挖掘过程中生成的非频繁项目集，有效降低主存压力，减少对数据库的扫描次数，能够对时变数据库进行高效地关联规则挖掘。通过实验证明了该算法能够有效地进行关联规则的挖掘，并在效率上有较大提高。

关键词: 关联规则, 增量式挖掘, 滑动窗口, 过滤器, 频繁项目集

Abstract: With the continual change and update of data in database, and the character of interesting in recent data in real life, an incremental mining algorithm of association rules based on sliding window filter(SWFAI) is presented. In the executing process of SWFAI algorithm, non-frequent item sets are given up in time by the way of dividing groups. The stress of main memory is abated, the times of scan of database are cut down, and the algorithm executes more efficient mining of association rules in time-variant database. An experiment is designed to prove that SWFAI algorithm can perform the mining of association rules availably, and the efficiency is improved at a certain extent.

Key words: Associational rules, Incremental mining, Sliding window, Filter, Frequent item sets