EcoSta 2017: Start Registration
View Submission - EcoSta2017
A0227
Title: An investigation of Ochi's estimator Authors:  Xiaoling Dou - Waseda University (Japan) [presenting]
Abstract: Ochi's estimator is proposed to estimate the autoregressive coefficient of the first order of autoregression (AR(1)) model by using two constants for the end points of the process. Classical estimators, such as the least square estimator, Burg's estimator and Yule-Walker's estimator of the parameter in AR(1) model are special choices of the constants in Ochi's estimator. First, we provide a simulation for AR(1) model and examine the performance of Ochi's estimator. Writing the autoregressive conditional heteroskedasticity model of order 1, ARCH(1), into a similar form to AR(1), we extend Ochi's estimator to the ARCH(1) model, introduce the ideas of the least squares estimator, Burg's estimator and Yule-Walker's estimator and compare the relationships of them with Ochi's estimator for ARCH(1) model. With a simulation, we investigate Ochi's estimator for ARCH(1) with different values of parameters and different sample sizes.