R. Jackson posted on Thursday, September 09, 2010 - 4:21 am
I have a question about mixture modeling in MPlus. I am looking for examples to do an analysis with unknown classes. In other words, I want MPlus to suggest the number of classes. Is this possible? The only examples I see are with known classes.
No, you must state the number of classes that you want the model to find. You can start with one or two classes and increase in subsequent analysis. Your choice of classes is guided by statistics and theory. See the Topic 5 video and course handout on the website.
R. Jackson posted on Thursday, September 09, 2010 - 3:59 pm
Thank you for the response. I am doing a simulation study on different "clustering" methods. I want to see which method does the best job of finding the correct number of clusters/classes. It looks like for mixture modeling I will have to run the model a few times with different classes and use the fit statistics to determine the "best" number of classes.