Thank you for your presentation and contribution to the conference. Can we conclude that CP and GCV are the best criteria for SETAR models? In the real world, we can not know the real data generation processes so we can not make a choice between CP and GCV? Do you consider to see Hannan Quinn criteria performance on SETAR models?
Thank you for your attention.
It is correct, we can conclude that Cp and GCV are the best selection methods.
In the context of real data processes, we will work on the Hannan Quinn criteria.
It is a further study that we plan to do I think.