Prev: Manually changing variable while running a loop
Next: Test the cexecution time of task on VxWorks
From: Noel Lu on 26 Apr 2010 04:20
Dear All!

When I use Oxford MFE toolbox to estimate parameters of the ARMA-GARCH model , I met a problem, and hope all your help!
When using MFE GARCH toolbox to estimate the parameters in the ARMA-GARCH model, I follow tow steps: 1. Using log-returns to estimate the parameters in the ARMA model and get residuals. 2. Using the residuals above, I fit the GARCH model. The question is, when I change the distribution hypothesis of residuals in the GARCH, such as from Gaussian to student T, the parameters’ estimators in the GARCH model will change, but the parameters’ estimators in the ARMA model don’t change. When I use the original GARCH toolbox in the MATLAB, all of them will change. What causes the difference, and how can I estimate the parameters in ARMA model using MFE GARCH toolbox when the distribution hypothesis of residual in the GARCH model changes?

Regards!
 | 
Pages: 1
Prev: Manually changing variable while running a loop
Next: Test the cexecution time of task on VxWorks