Title :
Parameter Estimation in an Autoregression Model with Infinite Variance
Author :
Alexandr, Markov
Author_Institution :
Tomsk State Univ., Tomsk
Abstract :
A weighted least squares procedure is proposed for parameter estimation in an autoregression model of first order with infinite variance of the noise. It is assumed that the noise distribution function belongs to the stable domain of attraction with index alpha, 0 < alpha < 2. The proposed procedure is shown to have higher rate of convergence to true value of the parameter as compared with usual least squares estimate. The limit distribution for weighted least squares estimates has been derived. The results of numerical simulations are given.
Keywords :
autoregressive processes; convergence of numerical methods; least squares approximations; parameter estimation; regression analysis; statistical distributions; autoregression model; convergence rate; infinite noise variance; limit distribution; noise distribution function; parameter estimation; weighted least squares estimation procedure; Autoregressive processes; Convergence; Distribution functions; Least squares approximation; Least squares methods; Numerical simulation; Parameter estimation; Probability distribution; Random variables; Stochastic resonance;
Conference_Titel :
Innovative Computing Information and Control, 2008. ICICIC '08. 3rd International Conference on
Conference_Location :
Dalian, Liaoning
Print_ISBN :
978-0-7695-3161-8
Electronic_ISBN :
978-0-7695-3161-8
DOI :
10.1109/ICICIC.2008.414
