Title :
Properties of the asymptotic quasi-likelihood estimate of a linear model
Author :
Mvoi, Sifa ; Lin, Yaii-Xia ; Biondini, Riccardo
Author_Institution :
Dept. of Appl. Stat., Wollongong Univ., NSW, Australia
Abstract :
Consider a model yt=ft(θ)+Mt , 0⩽t⩽T where θ∈Θ is an unknown parameter, ft(θ) is a linear predictable process, M t is a martingale difference and the nature of E(Mt2/ℱt-1) is unknown. The procedure for estimating θ is outlined. Some statistical properties of the estimates of θ obtained by this method are given. The advantage of the application of this procedure in the estimation of the dimension quantity of a fractal process is revealed. Examples from simulated data are given
Keywords :
convergence of numerical methods; fractals; linear systems; maximum likelihood estimation; prediction theory; signal sampling; statistical analysis; asymptotic quasi-likelihood estimate; convergence; dimension quantity; fixed sample space; fractal process; linear model; linear predictable process; martingale difference; parameter estimation; simulated data; statistical properties; Digital signal processing; Equations; Fractals; Least squares methods;
Conference_Titel :
Digital Signal Processing Proceedings, 1997. DSP 97., 1997 13th International Conference on
Conference_Location :
Santorini
Print_ISBN :
0-7803-4137-6
DOI :
10.1109/ICDSP.1997.628450
