Multiple change-point detection for regression curves

被引:0
作者
Wang, Yunlong [1 ]
机构
[1] Dongbei Univ Finance & Econ, Sch Data Sci & Artificial Intelligence, Dalian, Peoples R China
来源
CANADIAN JOURNAL OF STATISTICS-REVUE CANADIENNE DE STATISTIQUE | 2024年 / 52卷 / 04期
关键词
Change-point detection; cross-validation; information criterion; nonparametric regression; R PACKAGE; NUMBER; SEGMENTATION; MODELS;
D O I
10.1002/cjs.11816
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
Nonparametric estimation of a regression curve becomes crucial when the underlying dependence structure between covariates and responses is not explicit. While existing literature has addressed single change-point estimation for regression curves, the problem of multiple change points remains unresolved. In an effort to bridge this gap, this article introduces a nonparametric estimator for multiple change points by minimizing a penalized weighted sum of squared residuals, presenting consistent results under mild conditions. Additionally, we propose a cross-validation-based procedure that possesses the advantage of being tuning-free. Our simulation results showcase the competitive performance of these new procedures when compared with state-of-the-art methods. As an illustration of their utility, we apply these procedures to a real dataset. L'estimation non param & eacute;trique d'une courbe de r & eacute;gression devient cruciale lorsque la structure de d & eacute;pendance sous-jacente entre les covariables et les r & eacute;ponses n'est pas explicite. Bien que la litt & eacute;rature existante ait abord & eacute; l'estimation des points de changement uniques pour les courbes de r & eacute;gression, le probl & egrave;me des points de changement multiples n'est toujours pas r & eacute;solu. Dans le but de combler cet & eacute;cart, cet article pr & eacute;sente un estimateur non param & eacute;trique pour plusieurs points de changement en minimisant une somme pond & eacute;r & eacute;e p & eacute;nalis & eacute;e des r & eacute;sidus au carr & eacute;, pr & eacute;sentant des r & eacute;sultats coh & eacute;rents dans des conditions faibles. De plus, nous proposons une proc & eacute;dure bas & eacute;e sur la validation crois & eacute;e qui poss & egrave;de l'avantage d'& ecirc;tre sans r & eacute;glage. Nos r & eacute;sultats de simulation mettent en & eacute;vidence les performances comp & eacute;titives de ces nouvelles proc & eacute;dures par rapport aux m & eacute;thodes de pointe. Pour illustrer leur utilit & eacute;, nous appliquons ces proc & eacute;dures & agrave; un ensemble de donn & eacute;es r & eacute;el.
引用
收藏
页数:22
相关论文
共 38 条
[1]  
Arlot S, 2019, J SFDS, V160, P1
[2]   Estimation of a change-point in the mean function of functional data [J].
Aue, Alexander ;
Gabrys, Robertas ;
Horvath, Lajos ;
Kokoszka, Piotr .
JOURNAL OF MULTIVARIATE ANALYSIS, 2009, 100 (10) :2254-2269
[4]   Narrowest-over-threshold detection of multiple change points and change-point-like features [J].
Baranowski, Rafal ;
Chen, Yining ;
Fryzlewicz, Piotr .
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY, 2019, 81 (03) :649-672
[5]   Testing for Smooth Structural Changes in Time Series Models via Nonparametric Regression [J].
Chen, Bin ;
Hong, Yongmiao .
ECONOMETRICA, 2012, 80 (03) :1157-1183
[6]   GRAPH-BASED CHANGE-POINT DETECTION [J].
Chen, Hao ;
Zhang, Nancy .
ANNALS OF STATISTICS, 2015, 43 (01) :139-176
[7]   VARIABLE BANDWIDTH AND LOCAL LINEAR-REGRESSION SMOOTHERS [J].
FAN, JQ ;
GIJBELS, I .
ANNALS OF STATISTICS, 1992, 20 (04) :2008-2036
[8]   Changepoint Detection in the Presence of Outliers [J].
Fearnhead, Paul ;
Rigaill, Guillem .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 2019, 114 (525) :169-183
[9]   Multiscale change point inference [J].
Frick, Klaus ;
Munk, Axel ;
Sieling, Hannes .
JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-STATISTICAL METHODOLOGY, 2014, 76 (03) :495-580
[10]   WILD BINARY SEGMENTATION FOR MULTIPLE CHANGE-POINT DETECTION [J].
Fryzlewicz, Piotr .
ANNALS OF STATISTICS, 2014, 42 (06) :2243-2281