Weighted Association Patterns Mining Algorithm and its Application to Alarm Correlation Analysis inCommunication Network

摘要：

Alarm correlation analysis plays nil important role in communication network fault management.Using association rule algorithms in alarm correlation systems is a research hotspot.For alarms have different severity levels,an algorithm cared WAP Miner based on WAP-tree is proposed in which summation-weighted support and ratio-weighted support are introduced to generate alarm weighted association patterns.This algorithm generates less candidate paRerns,and enables us to effectively reduce the computational cost.WAP-tree improves the structure of FP-tree,so at least one third of memory is saved.Experiments demonstrate WAP Miner is much faster than MINWAL(O),and has great promise in knowledge acquisition for alarm correlation systems.

作者: Jing-Yao Feng Xing-Ming Li

会议类型: 国际会议

会议名称: The International Conference Information Computing and Automation(2007国际信息计算与自动化会议)

会议地点: 成都

会议语种:英文

页码: 1217-1220

在线出版日期: 2007-12-19（万方平台首次上网日期，不代表论文的发表时间）