Model Selection in Threshold Models
Repository DOI
Change log
Authors
Kapetanios, George
Abstract
This paper considers information criteria as model evaluation tools for nonlinear threshold models. Results concerning the consistency of information criteria in selecting the lag order of linear autoregressive models are extended to nonlinear autoregressive threshold models. Extensive Monte Carlo evidence of the small sample performance of a number of criteria is presented.
Description
Keywords
Is Part Of
Publisher
Faculty of Economics