首页 | 本学科首页   官方微博 | 高级检索  
     

数据流中异常模式的提取与趋势监测
引用本文:宋国杰,唐世渭,杨冬青,王腾蛟.数据流中异常模式的提取与趋势监测[J].计算机研究与发展,2004,41(10):1754-1759.
作者姓名:宋国杰  唐世渭  杨冬青  王腾蛟
作者单位:北京大学信息科学技术学院,北京,100871
基金项目:国家“八六三”高技术研究发展计划基金项目数据库重大专项课题 ( 2 0 0 2AA4Z3 440 ),国家“九七三”重点基础研究发展规划基金项目 (G19990 3 2 70 5 )
摘    要:研究的重点是数据流环境中异常模式的提取与趋势监测.主要贡献包括:①提出了一个进行异常模式发现的度量框架——强度比率,为异常模式挖掘提供了度量标准;②在基于异常模式求取的基础上,提出了利用回归分析方法——最小二乘法进行异常模式趋势监测.实验结果表明,提出的异常模式度量和求取算法是合理的,提出的趋势监测方法是有效的、可行的.

关 键 词:数据流  模式挖掘  异常模式  趋势监测  数据流分析

Extraction and Trend Detection of Unusual Patterns over Data Streams
SONG Guo Jie,TANG Shi Wei,YANG Dong Qing,and WANG Teng Jiao.Extraction and Trend Detection of Unusual Patterns over Data Streams[J].Journal of Computer Research and Development,2004,41(10):1754-1759.
Authors:SONG Guo Jie  TANG Shi Wei  YANG Dong Qing  and WANG Teng Jiao
Abstract:In this paper, emphasis is put on a hot topic abnormal behavior analysis, named unusual pattern extraction and trend detection Contributions of this paper include: (1) A new framework ratio of feature strength for unusual pattern extraction is proposed, which provides a standard for unusual pattern mining; (2) Based on the extraction of unusual pattern, a method for trend detection of unusual pattern is also proposed by using least square regression method The performance study shows that the framework of unusual pattern measurement and algorithm is reasonable, and the proposed trend detection method is executable and efficient
Keywords:data stream  pattern extraction  unusual pattern  trend detection  data stream analysis  
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号