LCA: understanding settings PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
Message/Author
 Jessica Hale posted on Wednesday, July 06, 2011 - 1:59 pm
I am wanting to determine the number of clusters are appropriate for lab data I have on 1210 female samples. I made all the variables binary.

I am new to using Mplus and am confused on the settings.

The below code,

ANALYSIS: Type=mixture;
STARTS = 5000 20;
STITERATIONS = 40;
LRTSTARTS = 0 0 40 10;
LRTBOOTSTRAP = 1000;

produced this warning:

THE BEST LOGLIKELIHOOD VALUE WAS NOT REPLICATED IN 2955 OUT OF 5000
BOOTSTRAP DRAWS. THE P-VALUE MAY NOT BE TRUSTWORTHY DUE TO LOCAL
MAXIMA. INCREASE THE NUMBER OF RANDOM STARTS USING THE LRTSTARTS OPTION.

So I tried the below code:

ANALYSIS: Type=mixture;
STARTS = 5000 20;
STITERATIONS = 40;
LRTSTARTS = 0 0 40 10;
LRTBOOTSTRAP = 5000;

same warning as above.

Then I changed the LRTSTARTS settings:

ANALYSIS: Type=mixture;
STARTS = 5000 20;
STITERATIONS = 40;
LRTSTARTS = 0 0 80 20;
LRTBOOTSTRAP = 5000;

and it ran fine.

I do not completely understand how to change the settings of LRTSTARTS and LRTBOOTSTRAP and could not find a clear definition of what each step is doing in the manual, only examples.

Thank you for any help that is given. It is greatly appreciated.
 Bengt O. Muthen posted on Wednesday, July 06, 2011 - 2:41 pm
It is the LRTSTARTS option that is the key here - see page 552 in the UG.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: