Dear Eviews,

I have different forecasts (of exchange rates) which I want to combine by means of Bayesian model averaging. I do not manage to program this in Eviews, do you have an example program for me?

Thank you very much!

Ralph

**Moderators:** EViews Gareth, EViews Jason, EViews Moderator

4 posts
• Page **1** of **1**

Dear Eviews,

I have different forecasts (of exchange rates) which I want to combine by means of Bayesian model averaging. I do not manage to program this in Eviews, do you have an example program for me?

Thank you very much!

Ralph

I have different forecasts (of exchange rates) which I want to combine by means of Bayesian model averaging. I do not manage to program this in Eviews, do you have an example program for me?

Thank you very much!

Ralph

- RalphVerhoeks
**Posts:**5**Joined:**Wed Feb 22, 2012 8:25 am

Bayesian Model Averaging is not built into EViews, but it depends what you want to do. Do you just want point estimates of forecasts, or do you need entire distributions? Have you already calculated the model probabilities?

- startz
- Non-normality and collinearity are NOT problems!
**Posts:**2427**Joined:**Wed Sep 17, 2008 2:25 pm

I just want to have point forecasts indeed, not the whole distributions. How should I calculate the model probabilities in Eviews? And then combine the forecasts in a Bayesian way?

Thanks again in advance!

Thanks again in advance!

- RalphVerhoeks
**Posts:**5**Joined:**Wed Feb 22, 2012 8:25 am

EViews isn't going to directly calculate the model probabilities. But you can use the Schwarz criterion to compute a reasonable large sample approximation for the posterior model odds. Then you multiply the forecast from each model by the model probability.

My favorite reference for this sort of stuff is Bayesian Econometrics, by Gary Koop.

My favorite reference for this sort of stuff is Bayesian Econometrics, by Gary Koop.

- startz
- Non-normality and collinearity are NOT problems!
**Posts:**2427**Joined:**Wed Sep 17, 2008 2:25 pm

4 posts
• Page **1** of **1**

Users browsing this forum: Google [Bot] and 2 guests