详细信息
文献类型:期刊文献
中文题名:会话流中Top-k闭序列模式的挖掘
英文题名:Top-k Closed Sequential Pattern Mining in Session Streams
作者:彭慧丽[1];张啸剑[2]
第一作者:彭慧丽
机构:[1]河南省直广播电视大学教务科;[2]河南财经学院计算机系
第一机构:河南省直广播电视大学教务科,郑州450008
年份:2009
卷号:35
期号:19
起止页码:86-87
中文期刊名:计算机工程
外文期刊名:Computer Engineering
收录:CSTPCD;;Scopus;北大核心:【北大核心2008】;CSCD:【CSCD2011_2012】;
基金:河南省科技厅基金资助项目"非线性降维技术在商业智能中的应用"(082300410110)
语种:中文
中文关键词:Top—k闭序列模式;加权调和平均数;调节因子
外文关键词:Top-k Closed Sequential Pattern(Topk_CSP); Weighted Harmonic Average(WHA); regulatory factor
摘要:在会话流中挖掘Top-k闭序列模式,存在因相关比率ρ的大小而导致的内存消耗和挖掘精度之间的冲突。基于False-Negative方法,提出Tstream算法,制定2种约束策略限制ρ。基于该策略设计加权调和计数函数,渐进计算每个模式的支持度。实验结果证明了该算法的有效性。
The current methods in session streams for mining Top-k Closed Sequential Pattern(Topk_CSP) may lead to a conflict between output precision and memory consumption because of using p. This paper proposes TStream algorithm, which is based on False-Negative approach. TStream utilizes two constraint strategies to restrict ρ, and employs a weighted harmonic count function to calculate the support of each pattern progressively. Experimental results show that the algorithm is efficient.
参考文献:
正在载入数据...