摘要
已有的滑动窗口数据流模型没有考虑过时数据和事务数量对挖掘结果的影响.针对该问题.提出了一种新的动态权值滑动窗口的数据流模型,并将该模型应用于数据流频繁项集挖掘中,设计了动态权值滑动窗口的频繁项集挖掘算法FIMDWS和改进算法FIMDWSW-Imp.通过实验对算法做了分析和评价.
In the existing literatures,sliding window models does not consider the mining effect of outdated data and the number of transactions. To solve this problem,this paper proposes a novel dynamic weighted sliding window data stream model. Firstly,we apply this model to the data stream frequent itemsets mining,and design and presents the algorithms of FIMDWSW( Frequent Itemsets Mining in Dynamic Weighted Sliding Window) and imporved FIMDWSW-IMP( Frequent Itemsets Mining in Dynamic Weighted Sliding Window-Improvment) for dynamic weighted sliding window data stream model. The performance of algorithms are analyzed and evaluated in numerical experiments.
引文
[1]Han J,Kamber M,Pei J.数据挖掘概念与技术[M].范明,孟小峰,译,北京:机械工业出版社,2012.
[2]寇香霞,任永功,宋奎勇.一种基于滑动窗口的数据流频繁项集挖掘算法[J].计算机应用与软件,2013,30(01):143-146.
[3]刘慧婷,周开申,赵鹏.基于滑动窗口的不确定性数据流频繁项集挖掘算法[J].计算机应用与软件,2015,32(12):12-15.
[4]Li H,Wang L.A Variable Size Sliding Window based Frequent Itemsets Mining Algorithm in Data Stream[C].International Conference on Materials Science.AIP Publishing LLC,2017:143-146.
[5]Farhat A,Gouider M.New Algorithm for Frequent Itemsets Mining from Evidential Data Streams[J].Procedia Computer Science,2016,96(11):645-653.