Predicted R-squared

[Jim Nieb]

Anyone know of a Stata command or user-written routine that generates the predicted R-squared after an estimation using -reg? Thanks.

[Clyde Schechter]

Predicted R-squared? What's that?

If you want the R squared from the regression, you'll find it in e(r2). If you want the adjuted R-squared, that's in e(r2_a).

[Jim Nieb]

Thanks for responding, Clyde. Please follow this link for a description of the predicted R-squared (sorry for long URL). Apparently, Minitab spits it out. https://blog.minitab.com/blog/advent...20observations.

[Felix Bittmann]

Apparently Maarten has provided a nice program for this kind of statistic in the past, see https://stackoverflow.com/questions/...-automatically

[Jim Nieb]

Great! Thanks, Felix.

[Clyde Schechter]

Thanks for the link in #3. Apparently it's a form of leave-out-one cross-validation.