燕山大学学报2011，Vol.35Issue(2)：115-120,6.DOI:10.3969/j.issn.1007-791X.2011.02.005
一种基于频繁模式有向无环图的数据流频繁模式挖掘算法
An algorithm based on frequent patterns directed acyclic graph for mining frequent patterns from data stream
摘要
Abstract
The algorithms based on FP-growth for mining frequent pattems need to scan the transaction database twice and give the threshold in advance, it also can not support time sensitive data stream.In this paper an algorithm based on frequent pattem directed acyclic graph far mining frequent pattems from data stream is proposed.Each transaction is given a number according to its coming time.Items contained in a transaction are sorted by their order, and the directed acyclic graph follows that order.A frequent pattem directed acyclic graph records the relation between transactions and items by its number.The process of pattem growth is to increase the transaction number of the directed edges.The atgonthm travels edges conversely by the same transaction number and products the conditional pattem bases, it can abstract the information of the conditional pattern bases according to dynamic threshold and scan the database only once to gain frequent patterns.The experiment shows that the proposed algorithm is better than FP-growth in execute time and the number of the storaged node is significiantly reduced.关键词
数据流/频繁模式/频繁模式有向无环图Key words
data stream/ frequent pattern/ frequent pattems directed acyclic graph分类
信息技术与安全科学引用本文复制引用
任家东,王倩,王蒙..一种基于频繁模式有向无环图的数据流频繁模式挖掘算法[J].燕山大学学报,2011,35(2):115-120,6.基金项目
河北省自然科学基金资助项目(F2008000888) （F2008000888）
