Bayesian nonstationary autoregressive models for biomedical signal analysis

被引：39

作者：

Cassidy, MJ

Penny, WD

机构：

[1] UCL, Sobell Dept Neurophysiol, Inst Neurol, London, England

[2] UCL, Wellcome Dept Imaging Neurosci, Inst Neurol, London, England

来源：

IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING | 2002年 / 49卷 / 10期

基金：

英国医学研究理事会; 英国惠康基金;

关键词：

autoregressive modeling; Bayesian; Kalman smoother; variational Bayes;

D O I：

10.1109/TBME.2002.803511

中图分类号：

R318 [生物医学工程];

学科分类号：

0831 ;

摘要：

We describe a variational Bayesian algorithm for the estimation of a multivariate autoregressive model with time-varying coefficients that adapt according to a linear dynamical system. The algorithm allows for time and frequency domain characterization of nonstationary multivariate signals and is especially suited to the analysis of event-related data. Results are presented on synthetic data and real electroencephalogram data recorded in event-related desynchronization and photic synchronization scenarios.

引用

页码：1142 / 1152

页数：11

共 29 条

[1] Adaptive AR modeling of nonstationary time series by means of Kalman filtering [J].