Bivariate change point detection: joint detection of changes in\n expectation and variance

2019/04/02 by Michael Messer, Messer, Michael · 1 citation
Mathematics · #Statistical Methods and Inference

paper · pdf · doi:10.48550/arxiv.1904.01320

Abstract

A method for change point detection is proposed. We consider a univariate\nsequence of independent random variables with piecewise constant expectation\nand variance, apart from which the distribution may vary periodically. We aim\nto detect change points in both expectation and variance. For that, we propose\na statistical test for the null hypothesis of no change points and an algorithm\nfor change point detection. Both are based on a bivariate moving sum approach\nthat jointly evaluates the mean and the empirical variance. The joint\nconsideration helps improve inference as compared to separate univariate\napproaches. We infer on the strength and the type of changes with confidence.\nNonparametric methodology supports the analysis of diverse data. Additionally,\na multi-scale approach addresses complex patterns in change points and effects.\nWe demonstrate the performance through theoretical results and simulation\nstudies. A companion R-package jcp (available on CRAN) is discussed.\n

Cited by

• Identification of Emotionally Stressful Periods Through Tracking Changes in Statistical Features of mHealth Data

Related

• A Note on Online Change Point Detection
• Asymptotic Distribution-Free Change-Point Detection for Multivariate and non-Euclidean Data
• Robust multiple change-point detection for multivariate variability using data depth
• Change point detection and inference in multivariable nonparametric models under mixing conditions
• A general procedure for change-point detection in multivariate time series