Abstract: Mining frequent itemsets is an important step in the association rules discovering.Mining the association rules has realistic meaning under the circumstance of the dynamic data changing.A new method for the mining of dynamic frequent itemsets is prensented.This method is developed based on previous episodes mining results.It only needs to scan part of the whole data set based on the previous results for the whole frequent itemsets mining at the end，and experimental results show that the performance of this algorithm is outperform the Apriori algorithm.

Key words: data mining, association rules mining, dymamic frequent itemset mining

摘要： 频繁项集挖掘是关联规则挖掘的重要步骤。在数据动态变化的环境下进行关联规则挖掘具有重要的现实意义。提出一种动态频繁项集挖掘算法，该算法建立在前一阶段挖掘的基础上，能避免过多地扫描数据库而影响挖掘性能，在最后生成全局频繁项集时，不需要全程扫描数据库，根据之前挖掘结果有选择地扫描相关的事务子集。实验表明，该算法挖掘性能远远优于Apriori算法，能有效地实现在数据动态变化环境下的挖掘频繁项集。

关键词: 数据挖掘, 关联规则挖掘, 动态频繁项集挖掘