Accurate ARMA signal identification-empirical results

Abstract

A recent ARMA (autoregressive moving average) algorithm for the identification of stochastic systems and signal models is discussed, and empirical results are given. The ARMA identifier discussed includes inherent order estimation. In comparison with other ARMA stochastic system identifiers, the new procedure has been empirically shown to yield highly accurate, unbiased results.<>