Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Stata extremely slow on bsqreg and sqreg command


From   Abekah Nkrumah <ankrumah@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Stata extremely slow on bsqreg and sqreg command
Date   Sun, 5 Aug 2012 18:29:10 +0100

Dear Stas,

Thanks very much for your response. I bet you I needed this wonderful
sense of humor as the model is still running as am writing. A good
Samaritan on this platform directed me to qreg2, unfortunately, that
seems to give some errors. What I plan doing is possibly to use qreg
which runs faster and prefix it with a bootstrap command to bootstrap
the standard errors. The only thing I will loose is the simultaneous
estimation. I think that I can give up that for some sleep. Thank you
very much.

On Sun, Aug 5, 2012 at 6:15 PM, Stas Kolenikov <skolenik@gmail.com> wrote:
> Dear fellow landscapers,
>
> I need fine sand to plant a few flowers. I put a few 5 cm rocks into
> my kitchen blender and started it on high grate. The blender seems to
> buzz and vibrate, but I don't see any sand coming out of it, just
> smoke.
>
> I bet your model is still around replication number 10 or so. One way
> to try to gauge how much time you need on your full data set is to
> figure out how the problem scales. Run it with 100, 200, 500 and 1000
> observations by 100, 200 and 400 replications (-set rmsg on- or use
> -timer-); fit a power model log( time ) = a + b log( sample size) + c(
> number of bootstrap replicates), and -predict- how things will look
> like for 100,000 observations. I did something of that kind timing my
> -confa- package in
> http://stata-journal.com/article.html?article=st0169 (may be that's
> something that the authors of computationally intensive packages
> should do). When people ask me about -confa- and show me their syntax
> with 50+ variables, I just [ palm face ] it.
>
> PS. I have MP4 -glamm- running for the 6th day doing multiple
> imputations on a model with two latent factors and a data set with 400
> observations. And another -glamm- problem with 350,000 observations
> running for the 12th day. Sometimes, you have to be patient.
>
> On Sun, Aug 5, 2012 at 3:47 AM, Abekah Nkrumah <ankrumah@gmail.com> wrote:
>> Dear Statalist,
>>
>> I will appreciate your help on this. I am running the sqreg and bsqreg
>> command with 400 replications. I executed the command around 23:30 hrs
>> Saturday and it is now 09:43 hrs Sunday and the command is still
>> running. Please note: I have 160,000 observations (of course some of
>> the observations will drop due to differences in observations per
>> variable). I am using  Stata SE version 11 running on a Macbook Pro
>> with 2.26 GHZ Intel core 2 duo and memory 2GB
>>
>> Is there anything I can do to get the command to run faster
>
>
> --
> ---- Stas Kolenikov
> -- http://stas.kolenikov.name
> ---- Senior Survey Statistician, Abt SRBI
> -- Opinions stated in this email are mine only, and do not reflect the
> position of my employer
> *
> *   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/



-- 
**********************************************
Dept. of Public Admin & Health Serv. Mgt
University of Ghana Business School
P.O. Box LG 78
Legon-Accra
Ghana
Tel: ++ 233 21 500159 Ext. 6247
       ++ 233 21 502258 Ext. 6247
       ++ 233 21 502255 Ext. 6247
Cell: ++233 243 198 313

Email: gankrumah@ug.edu.gh
           ankrumah@gmail.com
*
*   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/


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