Fast adaptive algorithms for abrupt change detection

被引:0
|
作者
Daniel Nikovski
Ankur Jain
机构
[1] Mitsubishi Electric Research Laboratories,
来源
Machine Learning | 2010年 / 79卷
关键词
Event detection; Distribution monitoring; CUSUM;
D O I
暂无
中图分类号
学科分类号
摘要
We propose two fast algorithms for abrupt change detection in streaming data that can operate on arbitrary unknown data distributions before and after the change. The first algorithm, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$\textsf{MB-GT}$\end{document} , computes efficiently the average Euclidean distance between all pairs of data points before and after the hypothesized change. The second algorithm, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$\textsf{MB-CUSUM}$\end{document} , computes the log-likelihood ratio statistic for the data distributions before and after the change, similarly to the classical CUSUM algorithm, but unlike that algorithm, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$\textsf{MB-CUSUM}$\end{document} does not need to know the exact distributions, and uses kernel density estimates instead. Although a straightforward computation of the two change statistics would have computational complexity of O(N4) with respect to the size N of the streaming data buffer, the proposed algorithms are able to use the computational structure of these statistics to achieve a computational complexity of only O(N2) and memory requirement of O(N). Furthermore, the algorithms perform surprisingly well on dependent observations generated by underlying dynamical systems, unlike traditional change detection algorithms.
引用
收藏
页码:283 / 306
页数:23
相关论文
共 50 条
  • [21] Nonparametric volatility change detection
    Mohr, Maria
    Neumeyer, Natalie
    SCANDINAVIAN JOURNAL OF STATISTICS, 2021, 48 (02) : 529 - 548
  • [22] Detection of Fast Transient Events in a Noisy Background
    Esman, Daniel J.
    Ataie, Vahid
    Kuo, Bill P. -P.
    Temprana, Eduardo
    Alic, Nikola
    Radic, Stojan
    JOURNAL OF LIGHTWAVE TECHNOLOGY, 2016, 34 (24) : 5669 - 5674
  • [23] Fast subset scan for multivariate event detection
    Neill, Daniel B.
    McFowland, Edward, III
    Zheng, Huanian
    STATISTICS IN MEDICINE, 2013, 32 (13) : 2185 - 2208
  • [24] Sequential change diagnosis revisited and the Adaptive Matrix CuSum
    Warner, Austin
    Fellouris, Georgios
    BERNOULLI, 2024, 30 (03) : 2228 - 2252
  • [25] Efficient Byzantine Sequential Change Detection
    Fellouris, Georgios
    Bayraktar, Erhan
    Lai, Lifeng
    IEEE TRANSACTIONS ON INFORMATION THEORY, 2018, 64 (05) : 3346 - 3360
  • [26] Quickest Change Detection With Observation Scheduling
    Ren, Xiaoqiang
    Johansson, Karl H.
    Shi, Ling
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2017, 62 (06) : 2635 - 2647
  • [27] Fast subset scan for spatial pattern detection
    Neill, Daniel B.
    JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY, 2012, 74 : 337 - 360
  • [28] Delay time in sequential detection of change
    Aue, A
    Horváth, L
    STATISTICS & PROBABILITY LETTERS, 2004, 67 (03) : 221 - 231
  • [29] A note on online change point detection
    Yu, Yi
    Padilla, Oscar Hernan Madrid
    Wang, Daren
    Rinaldo, Alessandro
    SEQUENTIAL ANALYSIS-DESIGN METHODS AND APPLICATIONS, 2023, 42 (04): : 438 - 471
  • [30] A Means for Tuning Primary Frequency Event Detection Algorithms
    Keene, Sean
    Hanks, Landon
    Bass, Robert B.
    2022 IEEE CONFERENCE ON TECHNOLOGIES FOR SUSTAINABILITY (SUSTECH), 2022, : 108 - 113