Convergence of least squares parameter estimates of weakly stationary time series models driven by uncorrelated processes
- 1 January 1977
- journal article
- research article
- Published by Taylor & Francis in International Journal of Systems Science
- Vol. 8 (1) , 25-30
- https://doi.org/10.1080/00207727708942020
Abstract
Convergence proofs for least squares identification of weakly stationary processes have been published by several researches. The best known is that of Mann and Wald (1943) where an independent and identically distributed (i.i.d.) input is considered, and where p-lim convergence is proved. Other proofs (Saridis and Stein 1908 also assume i.i.d. driving noise, though mean squares convergence is proved. In this paper no i.i.d. assumption is made but the more general bounded fourth moment white driving noise cases are covered, and a very short convergence proof is outlined. Convergence is shown to be in probability (p-lim), though under certain constraints almost sure convergence is also proved.Keywords
This publication has 3 references indexed in Scilit:
- Stochastic approximation algorithms for linear discrete-time system identificationIEEE Transactions on Automatic Control, 1968
- An Approach to Time Series AnalysisThe Annals of Mathematical Statistics, 1961
- On the Statistical Treatment of Linear Stochastic Difference EquationsEconometrica, 1943