|
![](http://www.statmodel.com/discussion/icons/csc/1px.gif) |
ML Integration with Mixtures |
![Previous](http://www.statmodel.com/discussion/icons/csc/navgen_arrow_left.gif) ![](http://www.statmodel.com/discussion/icons/csc/1px.gif)
|
Message/Author |
|
|
Hello, According to pages 482-483 of the manual, ML estimation with ordinal dependent variables requires Algorithm=Integration for Type=General, but does not require it for Type=Mixture. 1. Why is that? 2. Is it ok to use Type=Mixture with a single class, to fit models with ordinal dependent variables without using integration? |
|
|
There is a special path in mixture where random effects of continuous latent variables are not estimated and therefore numerical integration is not required. If a model requires numerical integration, it will require it using GENERAL or MIXTURE. |
|
Back to top |
|
|