Change-point detection with deep learning:A review  

作　　者：Ruiyu XU Zheren SONG Jianguo WU Chao WANG Shiyu ZHOU 

机构地区：[1]Department of Industrial Engineering and Management,Peking University,Beijing 100871,China [2]Department of Industrial and Systems Engineering,University of Iowa,Iowa City,IA 52242,USA [3]Department of Industrial and Systems Engineering,University of Wisconsin-Madison,Madison,WI 53706,USA

出　　处：《Frontiers of Engineering Management》2025年第1期154-176,共23页工程管理前沿(英文版)

基　　金：National Natural Science Foundation of China(Grant Nos.NSFC-71932006 and NSFC-72171003).

摘　　要：Recent advances in deep learning have led to the creation of various methods for change-point detection(CPD).These methods enhance the ability of CPD tech-niques to handle complex,high-dimensional data,making them more adaptable and less dependent on strict assump-tions about data distributions.CPD methods have also demonstrated high accuracy and have been applied across various fields,including manufacturing,healthcare,activity monitoring,finance,and environmental monitoring.This review provides an overview of how these methods are applied,the data sets they use,and how their performance is evaluated.It also organizes techniques into supervised and unsupervised categories,citing key studies.Finally,we explore ongoing challenges and suggest directions for future research to improve interpretability,generalizability,and real-world implementation.

关 键 词：change-point detection deep learning supervised learning unsupervised learning time-series analysis 

分 类 号：TP1[自动化与计算机技术—控制理论与控制工程]