How can I speed up gllamm?
| Title |
|
Speeding up gllamm |
| Author |
Minjeong Jeon, University of California, Berkeley |
| Date |
July 2012 |
There are two strategies to speed up gllamm:
First, collapse your data and specify frequency weights. For instance, if you have several identical level-2 units,
using level-2 weights makes gllamm enormously faster than without the weights.
See FAQ on creating frequency weights for gllamm,
and FAQ on specifying frequency weights in gllamm.
Second, use better starting values. For instance, you can use estimates obtained from a simpler model as starting values
for your target model using the from() option. Starting values for the parameters that do
not appear in the simple model are set to zero by default.
By doing this, you can save iterations and consequently reduce the computation time in gllamm.
Examples and documentation
- Description of
from() option on p.25
and examples with the from() option
on p.37, p.41, p.82, p.92, p.100 of Rabe-Hesketh, S., Skrondal, A. and Pickles, A. (2004).
GLLAMM Manual.
U.C. Berkeley Division of Biostatistics Working Paper Series. Working Paper 160.
- Examples with the
from() option
on p.17 and p.39 of glamm companion
for Rabe-Hesketh, S. and Skrondal, A. (2012). Multilevel and
Longitudinal Modeling Using Stata (3rd Edition). Volume
I: Continuous Responses. College Station, TX: Stata Press.
- Examples with the
from() option on p.598, 609-613, and p.887
and in exercises 10.8 and 11.7 in the book
Rabe-Hesketh, S. and Skrondal, A. (2012).
Multilevel and Longitudinal Modeling Using Stata (Third Edition).
Volume II: Categorical Responses, Counts, and Survival.
College Station, TX: Stata Press.
References
|