Search found 43 matches

Mon Apr 23, 2018 7:51 pm
Forum: MLwiN user forum
Topic: MCMC Macro
Replies: 22
Views: 14295

Re: MCMC Macro

Thank you so much for all your helpful explanations! Re your last point, I've got these covariates: age (continuous), education (categorical), household income (categorical), urbanicity (categorical), ... Of these, I realized that household income only varied at the level 2 (household-level) or abov...
Mon Apr 23, 2018 3:43 pm
Forum: MLwiN user forum
Topic: MCMC Macro
Replies: 22
Views: 14295

Re: MCMC Macro

This is really helpful and thank you again for the details! (1) Re the random part estimates for the region (highest level), I checked the data and it appears that the region (n=16) is sorted correctly. It's min=1, max=16, and categorical=true. In the view/edit data window, I can see that obs are so...
Mon Apr 23, 2018 1:47 pm
Forum: MLwiN user forum
Topic: MCMC Macro
Replies: 22
Views: 14295

Re: MCMC Macro

Hi Chris, Thank you for your advice! I've checked the covariance matrix (in the null model where it works) and it seems the variance estimates were: by IGLS estimation: - the highest level (region): 0.000(0.000) - the second highest (neighborhood): 0.034 (0.016) - the third highest (family): 1.0009 ...
Mon Apr 23, 2018 1:14 pm
Forum: MLwiN user forum
Topic: MCMC Macro
Replies: 22
Views: 14295

Re: MCMC Macro

Thank you for your kind response, Chris! Then, I just tried to run regular Poisson model (four-level with MCMC, offset=1 for all observations). It seems that the null model works, but when I added covariates, I got error messages as this: - "covariance matrix is not positive-definite" - &q...
Sat Apr 21, 2018 3:02 am
Forum: MLwiN user forum
Topic: MCMC Macro
Replies: 22
Views: 14295

Re: MCMC Macro

Also, I've tried to run just poisson and negative binomial model with the above setting, just by replacing 1 (for normal distribution) with 3 (for poisson) and 8 (for negative binomial), but in both cases, the model automatically switched to normal linear model. Am wondering what additional procedur...
Sat Apr 21, 2018 3:00 am
Forum: MLwiN user forum
Topic: Zero-inflated negative binomial models?
Replies: 5
Views: 4949

Re: Zero-inflated negative binomial models?

This looks very helpful and thanks for the useful discussion everyone!
Sat Apr 21, 2018 2:38 am
Forum: MLwiN user forum
Topic: MCMC Macro
Replies: 22
Views: 14295

Re: MCMC Macro

Dear Chris, Thank you so much again for your wonderful support! It's been really helpful and I appreciate it a lot! While making my projects move forward, I've got a few quick questions. Briefly, my project is as follows: (1) outcome: depressive symptoms score (0-27), which almost looks like zero-in...
Wed Feb 21, 2018 2:51 pm
Forum: MLwiN user forum
Topic: outcome prediction in MCMC model
Replies: 2
Views: 3111

Re: outcome prediction in MCMC model

Tue Feb 20, 2018 8:36 pm
Forum: MLwiN user forum
Topic: outcome prediction in MCMC model
Replies: 2
Views: 3111

outcome prediction in MCMC model

Hello, Hope this finds you all well! I've found this forum really helpful and useful and thank you for all great supports! Just got a quick question re a prediction in MCMC model (and in general non-MCMC model too). For instance, I've built a four-level linear model with MCMC (non-specified prior me...
Thu Aug 31, 2017 1:28 pm
Forum: Stat-JR user forum
Topic: imputation model specification
Replies: 23
Views: 17645

Re: imputation model specification

Thank you very much for all your helps Chris! Will try more and get it back to you again!
Best wishes,
Yongjoo