Online Program

Return to main conference page

All Times EDT

Friday, June 5
Computational Statistics
Computational Statistics 2
Fri, Jun 5, 11:15 AM - 12:50 PM
TBD
 

On the Estimation Bias in First-Order Bifurcating Autoregressive Models (308394)

Presentation

*Tamer Elbayoumi, North Carolina A&T State University 
Sayed Mostafa, North Carolina A&T State University 

Keywords: Bifurcating, Autoregressive, Bootstrap, Binary tree

In this study, bootstrap bias correction techniques are applied to stationary first order bifurcating autoregressive (BAR) models. The proposed resampling procedures are implemented to correct the bias that found with the least squares estimators of the BAR(1) model. The effectiveness of proposed approaches is studied empirically through extensive Monte Carlo simulations. In our simulation, we compared the mean squares errors and adjusted mean squares errors for the single bootstrap bias-corrected estimators, double bootstrap bias-corrected estimators, and fast double bootstrap bias-corrected estimators. Inference about the corrected estimators is implemented to demonstrate the confidence interval coverage.