DocumentCode :
3589585
Title :
Stochastic unobserved component models for adaptive signal extraction and forecasting
Author :
Young, Peter C. ; Tych, Wlodek ; Pedregal, D.J.
Author_Institution :
Centre for Res. on Environ. Syst., Lancaster Univ., UK
fYear :
1998
Firstpage :
234
Lastpage :
243
Abstract :
This paper is concerned with adaptive, off-line signal processing and forecasting for nonstationary signals described by the unobserved component model yt=Tt+St+f(ut)+Nt +et, for et~N{O, σ2}, where yt is the observed time series, Tt is a trend or low-frequency component, St is a periodic component possibly exhibiting temporal changes in both amplitude and phase, f(ut) captures the influence of a vector of exogenous variables ut, if necessary including stochastic, nonlinear static or dynamic relationships, Nt is a stochastic perturbation component, and et is an irregular component, normally defined for analytical convenience as a normally distributed Gaussian sequence with zero mean value and variance σ2 (i.e. discrete-time white noise). In order to allow for nonstationarity in the time series yt, the various components, including T t, are characterised by stochastic time-varying parameters each of which is defined as a nonstationary stochastic variable. The paper describes a new and flexible approach to off-line signal processing based on a dynamic harmonic regression (DHR) model of the unobserved components (UC) type and illustrates its performance via a typical problem of audio signal restoration
Keywords :
adaptive signal processing; prediction theory; stochastic systems; time series; white noise; DHR; adaptive off-line signal processing; adaptive signal extraction; audio signal restoration; discrete-time white noise; dynamic harmonic regression; dynamic relationships; forecasting; irregular component; low-frequency component; nonlinear static relationships; nonstationarity; nonstationary signals; nonstationary stochastic variable; normally distributed Gaussian sequence; periodic component; stochastic perturbation component; stochastic relationships; stochastic time-varying parameters; stochastic unobserved component models; zero mean value; Adaptive signal processing; Analysis of variance; Predictive models; Signal analysis; Signal processing; Signal restoration; Stochastic processes; Stochastic resonance; Time series analysis; White noise;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Neural Networks for Signal Processing VIII, 1998. Proceedings of the 1998 IEEE Signal Processing Society Workshop
ISSN :
1089-3555
Print_ISBN :
0-7803-5060-X
Type :
conf
DOI :
10.1109/NNSP.1998.710653
Filename :
710653
Link To Document :
بازگشت