Hadj-Amar Beniamino, Rand Bärbel Finkenstädt, Fiecas Mark, Lévi Francis, Huckstepp Robert
Department of Statistics, University of Warwick, Coventry, UK.
Division of Biostatistics, School of Public Health, University of Minnesota, Minneapolis, MN.
J Am Stat Assoc. 2019 Jul 9;115(531):1320-1335. doi: 10.1080/01621459.2019.1623043.
We propose a novel Bayesian methodology for analyzing nonstationary time series that exhibit oscillatory behavior. We approximate the time series using a piecewise oscillatory model with unknown periodicities, where our goal is to estimate the change-points while simultaneously identifying the potentially changing periodicities in the data. Our proposed methodology is based on a trans-dimensional Markov chain Monte Carlo algorithm that simultaneously updates the change-points and the periodicities relevant to any segment between them. We show that the proposed methodology successfully identifies time changing oscillatory behavior in two applications which are relevant to e-Health and sleep research, namely the occurrence of ultradian oscillations in human skin temperature during the time of night rest, and the detection of instances of sleep apnea in plethysmographic respiratory traces. Supplementary materials for this article are available online.
我们提出了一种新颖的贝叶斯方法来分析表现出振荡行为的非平稳时间序列。我们使用具有未知周期的分段振荡模型来近似时间序列，我们的目标是估计变化点，同时识别数据中潜在变化的周期。我们提出的方法基于一种跨维马尔可夫链蒙特卡罗算法，该算法同时更新变化点以及与它们之间任何段相关的周期。我们表明，所提出的方法在与电子健康和睡眠研究相关的两个应用中成功识别了随时间变化的振荡行为，即夜间休息期间人体皮肤温度的超日振荡的发生，以及在体积描记呼吸轨迹中检测睡眠呼吸暂停的实例。本文的补充材料可在线获取。
