基于数据流频繁闭合模式的挖掘

用挖掘频繁闭合模式集代替挖掘频繁模式集是近年来提出的一个重要策略。根据数据流的特点，提出了一种基于滑动窗口的频繁闭合模式的新方法DSFC_Mine。该算法以滑动窗口中的基本窗口为更新单位，利用改进的CHARM算法计算每个基本窗口的潜在频繁闭合项集，将它们存储到一种新的数据结构中，利用该数据结构可以快速地挖掘滑动窗口中的所有频繁闭合项集。实验验证了该算法在时间上和空间上的可行性和有效性。

Mining frequent closed patterns over data stream

Recently, frequent closed patterns mining has been an important method to replace the frequent patterns mining. According to the features of data stream, a new algorithm called DSFC_Mine was proposed to solve the problem of mining the frequent closed patterns from sliding window. The basic window of a sliding window was served as an updating unit in this algorithm. And all potential frequent closed patterns of every basic window were mined by the improved CHARM algorithm. Those patterns were stored in a new data structure. And the frequent closed patterns in a sliding window could be rapidly found based on the new data structure. The experimental result shows the feasibility and effectiveness of the algorithm.