Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Nick Cox <njcoxstata@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: RE: How to generate binary y? |
Date | Fri, 4 May 2012 07:06:18 +0100 |
I think most people round to the nearest integer, so -round(predicted,1)- rounds 0.5 up to 1 and lower values to 0. There is no need to round unless you want to impute or predict at individual level. Nick On Thu, May 3, 2012 at 11:04 PM, Sumedha Gupta <suguptastatlist@gmail.com> wrote: > Thank you so much for your response. After I predict out of sample how > do I allocate 0/1 to the continuous predicted value i.e. what is the > cut off? Probably this is a very basic question but I will really > appreciate your help. On Thu, May 3, 2012 at 3:49 PM, Nick Cox <n.j.cox@durham.ac.uk> wrote: >> Sounds as if you just need to use -predict- after -probit-. -predict- will happily predict out-of-sample so long as all predictors in the model are non-missing. Sumedha Gupta >> I have 2 samples one in which both x and y are observed and the other >> in which only the x's are observed. Y is binary and some of the x's >> are binary as well. I ran a probit of y on x for sample 1 and got all >> the betas. Now using these beta's would it be possible to generate y >> in sample 2 as well? This would be easy to do in the linear case... >> but for a probit is there a command to do that? * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/