In and Out of Sample Estimation/Tests

[5feira]

I'm really new to Eviews, so sorry if I'm asking a too dumb question.
This is what I need:
I have two series (variable X and Y) which are correlated as Y=aX+b.
I need to test 3/4 of the series with different models (all of the possible ones Eviews let me) to find constants "a" and "b". And then test this constants with the out-of-sample data to check which models fits best this out-of-sample data.
Could anyone please help me?

[EViews Gareth]

There's nothing built in that will do this.

To be honest, it seems like a strange thing to want to do.

[5feira]

OK. Thanks!
One more question...
Is there a way I could estimate this constants a and b with, let's say EGARCH model, and then check the statistics for the out-of-sample data?
Example:
I have two series of data (X and Y) from 2000 to 2014. Then I would like to estimate constants "a" and "b" using data from 2000 to 2010 and then check the error of the forecasts for 2011 to 2014 (using these "a" and "b") and the actual data.

Thanks, Gareth!

[EViews Gareth]

Yes, create an equation, estimate with the sample 2000 2010, then use the forecast button to forecast. EViews will tell you the forecast evaluation statistics.

[5feira]

Thanks a lot, Gareth!