Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: initial values for -xtmelogit-?


From   Jeph Herrin <junk@spandrel.net>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: initial values for -xtmelogit-?
Date   Tue, 08 Apr 2008 09:29:35 -0400

Thanks, -from()- isn't documented but Bob Gutierrez pointed this out
to me off list.

cheers,
Jeph



Austin Nichols wrote:
webuse bangladesh, clear
xtmelogit c_use urban age child* || district:
mat b=e(b)
xtmelogit c_use urban age child* || district: , from(b)
xtmelogit c_use urban age child* || district: , from(b) refineopts(iterate(0))

On Mon, Apr 7, 2008 at 5:53 PM, Jeph Herrin <junk@spandrel.net> wrote:
 I am doing a simulation that requires estimating
 a model using -xtmelogit- for each of 1000 iterations.

 Since the same model is being estimated using samples
 drawn each time from the same dataset, it seems that
 there should be some way to use the results of each
 iteration as initial values in the next. However,
 -xtmelogit- doesn't seem to accept -initial()- as
 a maximize option. Is there a way I don't see to
 persuade -xtmelogit- to start in a certain place?
 Getting the initial values seems to be the most time
 consuming bit of the estimation.

 (As an aside, I could use -xtlogit, re-, which is considerably
 faster, if only -predict- could be persuaded to predict
 random effects, something which apparently is just a minor
 fix:

 http://www.stata.com/statalist/archive/2007-03/msg00135.html

 I had hoped that Stata 10 would fix this, but apparently -xtlogit-
 is on the path to deprecation.)

 Jeph
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index