|
![](http://www.statmodel.com/discussion/icons/csc/1px.gif) |
Lots of latent variables with varianc... |
![Previous](http://www.statmodel.com/discussion/icons/csc/navgen_arrow_left.gif) ![](http://www.statmodel.com/discussion/icons/csc/1px.gif)
|
Message/Author |
|
genwei posted on Wednesday, September 09, 2015 - 12:18 am
|
|
|
Hi Professors, EXAMPLE 7.26 showed how to model a latent variables with variance zero, and it is easy to extend it to multiple latent variables case. However, in my model, I get 18 such latent variables (or even more). Running mplus will tremendously consume memory (over 8GB) and extremely slow (ALGORITHM=INTEGRATION; and Monte Carlo seems will not solve the problem). I also tried BAYES method, but it seems not working for my model including ON command yet. My question is any alternative way to overcome this estimation problem? |
|
genwei posted on Wednesday, September 09, 2015 - 12:22 am
|
|
|
And, I prefer to use ML method rather than WLSMV because of nonignorable missing data existing (simulation data). |
|
|
I don't this problem can be done. |
|
Back to top |
|
|