Last 3 Days
BLRT problems with neg-bin GMM
Latent Variable Mixture Modeling
posted on Tuesday, October 02, 2018 - 12:20 am
we are having some problems with the BLRT when fitting a negative-binomial growth mixture model - we obtain 100% non-convergence and the estimation is extremely slow.
Lo-Mendel Rubin is fine, and both bootstrap tests are fine when using Poisson instead of NB.
I'm new to modelling count data in this way and am wondering if there might be a sensible technical reason why I am having problems here. Any thoughts would be gratefully received.
best wishes, Jon
Bengt O. Muthen
posted on Tuesday, October 02, 2018 - 5:27 pm
We found a glitch in BLRT for this situation - it will be fixed in the forthcoming Mplus Version 8.2. Thanks for pointing this out.
posted on Wednesday, October 03, 2018 - 2:18 am
ahh great, thanks Bengt
Back to top
Add Your Message Here
This is a private posting area. Only registered users and moderators may post messages here.
Enable HTML code in message
Automatically activate URLs in message