Using penalized contrasts for the change-point problem
Using penalized contrasts for the change-point problem
复制标题
DOI：
10.1016/j.sigpro.2005.01.012
复制
发表时间：
2005-08-01
影响因子：
4.4
通讯作者：
Lavielle, M
中科院分区：
文献类型：
--
作者：
Lavielle, M
提出了一种基于惩罚对比度的模型选择方法。将该方法应用于变点问题，估计变点的数目及其位置。我们的目标是通过构建可用于不同实际情况的算法来完成以前的渐近结果。首先，我们提出了一种自适应选择罚函数来自动估计模型的维数，即，改变点的数量。在贝叶斯框架中，我们定义的后验分布的变点序列作为惩罚对比度的函数。MCMC程序可用于对该后验分布进行采样。该分布的参数估计与随机版本的EM算法（SAEM）。EEG分析和蒙特-卡罗实验的应用说明了这些算法。(c)2005 Elsevier B.V.保留所有权利。
A methodology for model selection based on a penalized contrast is developed. This methodology is applied to the change-point problem, for estimating the number of change points and their location. We aim to complete previous asymptotic results by constructing algorithms that can be used in diverse practical situations. First, we propose an adaptive choice of the penalty function for automatically estimating the dimension of the model, i.e., the number of change points. In a Bayesian framework, we define the posterior distribution of the change-point sequence as a function of the penalized contrast. MCMC procedures are available for sampling this posterior distribution. The parameters of this distribution are estimated with a stochastic version of EM algorithm (SAEM). An application to EEG analysis and some Monte-Carlo experiments illustrate these algorithms. (c) 2005 Elsevier B.V. All rights reserved.