首页> 外国专利> TIME SERIES DATA ABNORMALITY MONITORING DEVICE AND TIME SERIES DATA ABNORMALITY MONITORING METHOD

TIME SERIES DATA ABNORMALITY MONITORING DEVICE AND TIME SERIES DATA ABNORMALITY MONITORING METHOD

机译:时间序列数据异常监视装置及时间序列数据异常监视方法

摘要

PROBLEM TO BE SOLVED: To properly detect an abnormality due to unspecific variation of time series data of such as the communication traffic of a network device.SOLUTION: A time series data abnormality monitoring device 10 includes a calculation unit 11 for obtaining various variation patterns according to different conditions from the time series data by the communication traffic of a network device, and storing the various variation patterns in a variation pattern storage unit 17 while classifying, and an estimation unit 12 for obtaining an estimate, i.e., the value in the future when the time series data does not exist. A comparison unit 13 for acquiring various variation patterns in the past, corresponding to the zone of that estimate, from the storage unit, comparing each of the various variation patterns with the estimate, and obtaining multiple differences of these comparisons, and a selection unit 14 for selecting a variation pattern, corresponding to the smallest difference out of the multiple differences, as an adaptive model, are included furthermore. An abnormality detector 15 for detecting an abnormality when the time series data, corresponding to a variation pattern as its adaptive model, exceeds a predetermined threshold is provided.SELECTED DRAWING: Figure 1
机译:解决的问题:为了适当地检测由于诸如网络设备的通信业务之类的时间序列数据的非特定变化而导致的异常。解决方案:时间序列数据异常监视设备10包括计算单元11,用于根据通过网络设备的通信流量从时间序列数据中获取不同的条件,并在分类的同时将各种变化模式存储在变化模式存储单元17中,以及用于获得估计值(即将来的值)的估计单元12当时间序列数据不存在时。比较单元13,用于从存储单元中获取过去与该估计的区域相对应的各种变化模式,将各个变化模式与估计值进行比较,并获得这些比较的多个差,选择单元14此外,还包括用于选择与多个差异中的最小差异相对应的变化模式的自适应模型。设有异常检测器15,用于当与作为其自适应模型的变化模式相对应的时间序列数据超过预定阈值时检测异常。

著录项

  • 公开/公告号JP2017143399A

    专利类型

  • 公开/公告日2017-08-17

    原文格式PDF

  • 申请/专利权人 NIPPON TELEGR & TELEPH CORP NTT;

    申请/专利号JP20160023207

  • 发明设计人 TANJI NAOYUKI;MORIYA TAKAAKI;

    申请日2016-02-10

  • 分类号H04L12/70;

  • 国家 JP

  • 入库时间 2022-08-21 14:00:36

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号