how IS that possible? Unless you mean adjusted R^2
got an out-of-sample R^2 of -5.8 the other day somehow hadn't occurred to me that was possible but now I'm curious how far I can push it
-
-
-
Take 1-(RSS/TSS) Model fit is worse than a constant at the mean of the dependent variable Possible, because model fit on separate data
-
oh, right, out-of-sample. yeah that is actually impressively bad!
-
you can make it arbitrarily large. Estimate sample mean with N=1, choose out-of-sample N=1 arbitrarily far away
-
Or maybe N=2 so TSS isn't 0
-
yes I think for this to be fun it needs restrictions perhaps coming from *unintentionally* overfitting rather than deliberate
-
in that case you find yourself at a likely disadvantage compared to someone who doesn't know what overfitting is
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.