On Fri, 9 Jul 2010, Pindar wrote:
when performing a multinomial logit estimation Gretl gives as
fitted values a certain category for each id, which is nice.
However, is there a quick way to get the estimated probabilities
for each category and each id?
If I'm not mistaken, here's how you could get the probabilities
for the keane.inp example in the gretl package:
<script>
open keane.gdt
list reg = const educ exper expersq black
smpl (year=87 && ok(reg)) --restrict
logit status reg --multinomial
scalar k = nelem(reg)
matrix b2 = $coeff[1:k]
matrix b3 = $coeff[k+1:]
matrix X = {reg}
matrix Xb = exp(1~(X*b2)~(X*b3))
matrix Probs = Xb ./ sumr(Xb)
print Probs
</script>
By the way, the command forecast in the multinomial logit window
leads to a
crash of Gretl.
Oops, thanks the report. That is not properly hooked up, so I've
disabled it for the moment.
And for further development one could think of a "base"
option,
n'est pas?
Yes, probably.
Allin Cottrell