多数据流的频繁模式挖掘算法研究

Research on Mining Frequent Patterns over Multiple Data Streams

提出一种多数据流频繁模式挖掘算法Multiple Data Stream Mining(MDSM).MDSM算法用Multiple Frequent Pattern Tree(MFP-Tree)结构来存储多数据流中的频繁项集和潜在的频繁项集,并通过增量更新的方式高效地挖掘多数据流中的协同频繁模式和比较频繁模式.通过理论分析和实验证明其可行性.

This paper proposes an algorithm of mining frequent patterns over Multiple data streams：Multiple Data Stream Mining（MDSM）. The algorithm constructs a new Multiple Frequent Pattern Tree to store the historical frequent itemsets and the potential frequent itemsets of the multiple data streams. Furthermore,make use of the incremental updates to mine these itemsets collaborative and comparative pattern. Theoretical and empirical studies demonstrate the utility of the proposed algorithm.