out-of-sample testing of a regression

teho · Postby **teho** » Fri Feb 17, 2012 3:09 am

I'd like to test my regression model out-of-sample, so I've run an OLS regression for period 2000-2005. I'd like to test the coefficients from this equation in the dataset between 2005-2010. How could the coefficients be fixed in the new regression using data from 2005-2010, so that I'm able to know how well the estimated coefficients fit the data out-of-sample. Thanks in advance

EViews Glenn · Postby **EViews Glenn** » Fri Feb 17, 2012 2:52 pm

Look at the discussion of forecasting in the manual.

teho · Postby **teho** » Mon Feb 20, 2012 8:14 am

Is it possible to get t-statistics and R2 values for the forecasted model?

EViews Gareth · Postby **EViews Gareth** » Mon Feb 20, 2012 9:28 am

I'm not sure I follow.

teho · Postby **teho** » Mon Feb 20, 2012 11:48 am

I Was wondering if it's possible to calculate R^2 for the model and t-values for coefficients, even though they are fixed. have to survive with the statistics given, thanks!

EViews Gareth · Postby **EViews Gareth** » Mon Feb 20, 2012 2:17 pm

That doesn't make sense, econometrically.

EViews.com