基于滑动窗口的数据流最大频繁项集的挖掘

Mining maximal frequent itemsets in a sliding window over data streams

鉴于频繁项集存在数据和模式冗余的问题,挖掘数据流最大频繁项集的算法引起了极大的关注,本文提出了一种挖掘数据流滑动窗口内最大频繁项集算法——MMFI-SW算法。该算法首先使用类似FP-tree的数据结构记录最新到达的数据流信息,同时删除过时的数据和大量的不频繁项目,然后设计一个创新的方法有效地从数据流滑动窗口中输出最大频繁项集。理论分析与实验结果表明,MMFI-SW算法具有较低的时间复杂度。

In consideration of the problem of data and pattern redundancy in frequent itemsets mining and the close attention to the study of mining maximal frequent itemsets from data streams, this paper presents the MMFI-SW algorithm for mining maximal frequent itemsets in a sliding window over data streams. Firstly, it uses a data structure based on FP-tree to record the current information in streams, at the same time, the obsolete items and a lot of infrequent items are deleted by pruning the tree. Then, it designs a novel method to mine the set of all maximal frequent itemsets in a sliding window over data streams. The theoretical analysis and the experimental results show that the proposed method is efficient.