Title :
On estimating the orders of an autoregressive moving-average process with uncertain observations
Author_Institution :
M. I. T. Lincoln Laboratory, Lexington, MA, USA
fDate :
10/1/1972 12:00:00 AM
Abstract :
A method is proposed for estimating the orders of an autoregressive moving-average process solely from the observations of the noise-corrupted output. The order of the autoregressive (number of poles) can be determined by testing the singularity of the correlation matrix, while the order of the moving-average (number of zeros) can be determined from observing the repetitive pattern of the correlation equations.
Keywords :
Autoregressive moving-average processes; Estimation; Asymptotic stability; Differential equations; Notice of Violation; Poles and zeros; Stability criteria; Stochastic processes; Stochastic resonance; Stochastic systems; Sufficient conditions; Testing;
Journal_Title :
Automatic Control, IEEE Transactions on
DOI :
10.1109/TAC.1972.1100122