Combining pattern recognition techniques with Akaike´s information criteria for identifying ARMA models

Author

Wang, Liang ; Libert, Gaëtan A.

Author_Institution

Dept. of Electr. Eng., Texas A&M Univ., College Station, TX, USA

Volume

42

Issue

6

fYear

1994

fDate

6/1/1994 12:00:00 AM

Firstpage

1388

Lastpage

1396

Abstract

ARMA models are identified by combining pattern recognition techniques with Akaike´s (1974, 1979) information criteria. First, pattern vectors of ARMA models are obtained by the extended sample autocorrelation functions method proposed by Tsay and Tiao (1984). Second, decision functions of various training samples are specified by the perceptron algorithm used in learning machines. Third, Akaike´s AIC and BIC criteria are introduced. The practical utility of the proposed approach is illustrated by both simulated and practical data

Keywords

correlation methods; learning (artificial intelligence); neural nets; parameter estimation; pattern recognition; stochastic processes; time series; AIC criterion; ARMA models; Akaike´s information criteria; BIC criterion; decision functions; extended sample autocorrelation functions method; pattern recognition techniques; pattern vectors; perceptron algorithm; training samples; Autocorrelation; Computer science; Decision support systems; Equations; Measurement errors; Pattern matching; Pattern recognition; Polynomials; Predictive models; White noise;

fLanguage

English

Journal_Title

Signal Processing, IEEE Transactions on

Publisher

ieee

ISSN

1053-587X

Type

jour

DOI

10.1109/78.286955

Filename

286955