Am 23.08.2013 12:39, schrieb Artur Tarassow:
Hi,
I am just wondering if there already exists a built-in accessors for the
number of cases correctly predicted after a logit/probit estimation? I
couldn't find any in the manual, but maybe I missed something.
I haven't yet worked with these models in gretl, but assuming that $yhat
produces something useful there (like the probability), it should be
pretty easy, no?:
numcorrectpredict = sum(($yhat>0.5)==binarydep)
hth,
sven