I would like to run a multinomial logit estimate in python or R in which samples have different weights (these are population weights in a survey). statsmodels' MNLogit does not seem to provide this. Neither does mlogit in R. Is there another way to do it?
Asked
Active
Viewed 478 times
0
CPBL
- 3,783
- 4
- 34
- 44
-
Not available in statsmodels, but maybe just a few hours or days of work to add it. – Josef Dec 15 '16 at 22:03
-
Try `multinom()` in package `nnet`? – sirallen Dec 16 '16 at 06:00
-
@sirallen: I don't see this option there, but if you want to suggest how, please... – CPBL Dec 18 '16 at 04:01