摘要
首先回顾了电信网告警数据库中数据挖掘技术应用的研究进展,然后对告警模型进行了形式化描述,并阐述了告警模式挖掘算法W INEPI的基本思想,接着讨论了时间窗宽度改变情况下对候选集规模削减的两个约束条件,基于此提出了一种基于时间窗约束的增量式频繁情景挖掘算法。实验结果表明,该算法的执行效率在一定条件下比原有W INEPI算法有显著提高。
Firstly, the research progress on applying KDD techniques to telecommunication network databases is reviewed. Secondly, the model of alarms is descripted formally, and the basic idea of WINEPI which is the alarm patterns mining algorithm is elucidated. Then two constrained conditions to reduce the scale of candidate set when the width of a time window changes are discussed, based on which a time window constrained based incremental frequent episodes mining algorithm is proposed. Experimental results show that the algorithm is much more efficient than the previous WINEPI algorithm on proper conditions.
出处
《计算机应用研究》
CSCD
北大核心
2006年第3期257-260,共4页
Application Research of Computers
基金
教育部科学技术研究重点项目(02029)
关键词
数据挖掘
增量式挖掘
告警
频繁情景
时间窗
Data Mining
Incremental Mining
Alarm
Frequent Episodes
Time Window