滑动窗口中数据流频繁项集挖掘方法  

Algorithm for mining frequent itemsets from sliding window over data streams

作　　者：张月琴[1] 

机构地区：[1]南京工业大学电子与信息工程学院,南京210009

出　　处：《计算机工程与应用》2010年第16期132-134,共3页Computer Engineering and Applications

基　　金：南京工业大学青年教师学术基金(No.39709013)

摘　　要：根据数据流的流动性与连续性,提出了一种滑动窗口中频繁项集挖掘算法NSW,满足了人们快速获取最近到达数据中频繁项集的需求。该算法采用二进制矩阵表示滑动窗口中的事务列表,通过直接删除最老事务、不产生候选项集等方法控制时间和空间的开销。实验表明,该算法具有较好的时间和空间效率。According to the mobility and flowing of data streams,an algorithm called NSW is proposed to mine the frequent itemsets from a sliding window over data streams,and it meets the needs of people getting the frequent itemsets over datas arrived in recently.The binary matrix representation is adopted in the proposed algorithm to express the transaction list from a sliding window.The oldest transaction is deleted directly,and not generating the candidate itemsets but generating the frequent itemsets directly.These methods greatly control the space and time.The experiment results show that this algorithm has a good performance in speed and space.

关 键 词：数据挖掘 数据流 频繁项集 滑动窗口 矩阵 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]