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]

st: FW: Re: xtmixed slope cross validation


From   ymarchenko@stata.com (Yulia Marchenko, StataCorp LP)
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   st: FW: Re: xtmixed slope cross validation
Date   Fri, 02 Dec 2011 14:46:33 -0600

Petra <petroula.proitsi@kcl.ac.uk> has a follow-up question about computing
BLUPs after -xtmixed- using test data after the model is fit using training
data:

> however, I get all my people with empty values when i do this: my exact code
> is
>
>  . use:" data...."
>  * i have already a variable training =1 if it is training and 0 if it is test
>  . xi:xtmixed var1 days ....... if training==1 || id:days, reml cov(un)
>  . predict rs1 rs2 it training!=1, reffects
>  
> why do you think this is? by the way im using stata10 ic
> do you think this may be the reason?

In this case, the most likely reason for getting missing values of BLUPs after
-predict, reffects- is that the group identifiers in the test data are
different from the group identifiers in the training data.  When creating a
training indicator, Petra needs to make sure that each group is represented in
both test and training datasets.  For example,

  . webuse pig, clear
    // set seed for reproducibility
  . set seed 287634
    // select 50% of obs within each group into a training sample
  . by id, sort: gen training = runiform()>0.5
  . xtmixed weight week if training || id: week
  . predict r* if !training, reffects
  . summarize r*


-- Yulia
ymarchenko@stata.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