Hidden Markov model w/ 7 time points PreviousNext
Mplus Discussion > Categorical Data Modeling >
Message/Author
 Renate Houts posted on Tuesday, October 25, 2005 - 8:27 am
I've been trying to run a hidden Markov model with 7 time points in version 3.12. My outcomes are dichotomous and I specified measurement equality over time. Unlike Example 8.12, there are no constraints on the transition matrix.

The model appears to run fine. However, the output is incomplete (it stops after the title "ODDS RATIO RESULTS") and the program does not create my output file with CPROBABILITIES. When I delete the last time point, the model runs fine and the output is complete.

Any ideas on what might be going on?
 Linda K. Muthen posted on Tuesday, October 25, 2005 - 9:04 am
I suggest trying it with Version 3.13 and if the same thing happens sending the input, data, output, and your license number to support@statmodel.com.
 Blaze Aylmer posted on Wednesday, October 26, 2005 - 3:12 pm
Linda,Bengt

Does the approach of multinomial regression example for LTA apply to a mixed markov model with three classes and three predictor variables or is a different approach required in Mplus.

I'm using a five point likert scale to measure my DV and wondered if it would impact the size of the transition matrix? Does MPLUS have a feature to collapse a five point scale into a dichotomous one?

Blaze
 Linda K. Muthen posted on Wednesday, October 26, 2005 - 5:04 pm
In Mplus the regression of a categorical latent variable on a covariate is a multinomial logisitc regression.

The transition matrix is not affected by the number of categories in the dependent variables.

The DEFINE command has an option for dichotomizing variables. It is called CUT.
 Geoffrey Hammond posted on Tuesday, June 27, 2006 - 6:51 am
I have recently been applying Markov analysis to my data and have found the optimal model chosen is a two chain hiddden markov model with measurement error. How do i properly assess the transitional probabilites for each individual chain? The output only provides me with the the transition and stability probabilities for the population as a whole, not divided by chain. Do i derive this by cross-tabulation of the conditional probabilites output, stratified by chain? Is this correct?? Thanks in advance,

Geoff
 Bengt O. Muthen posted on Wednesday, June 28, 2006 - 2:27 am
This is correct.
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: