Title :
Two-step estimation of misspecified, non-Gaussian ARMA processes
Author :
Gerencsér, László ; Michaletzky, György ; Reppa, Zoltàn
Abstract :
ARMA modelling of economic time series leads to processes with heavy-tailed marginal distribution. We present methods of estimating the system and noise parameters of such processes. Asymptotic properties of the quasi maximum likelihood and partially adaptive estimates are discussed. The results are generalizations of the previous work, and are applicable to derive good approximation formulas for the error of adaptive predictors.
Keywords :
adaptive estimation; autoregressive moving average processes; probability; time series; economic time series; noise parameters; nonGaussian ARMA processes; partially adaptive estimates; quasi maximum likelihood; two-step estimation; Economic forecasting; Gaussian processes; Hydrology; Least squares approximation; Linear systems; Maximum likelihood estimation; Parameter estimation; Robustness; Stock markets; Technological innovation;
Conference_Titel :
Decision and Control, 2002, Proceedings of the 41st IEEE Conference on
Print_ISBN :
0-7803-7516-5
DOI :
10.1109/CDC.2002.1184789