If my research is to find an optimized method to analyze procrastination in online submissions among students, and I want to use Finite Mixture Model (FMM), can you suggest on which part in the algorithm can I focus on in order to make the modification as part of my research contribution?
Hajar,
can you be more specific on what you are trying to model? Are you planning to explore whether there are subgroups of students based on when they submit their work?
Heiko Breitsohl thank you for your interest. I am trying to see subgroups of students based on their submission duration, task difficulty levels, number of attempts, skills, and their grades.
Okay, do you expect that these groups will differ on the
- means
- variances
- covariances
of those variables?
Hi, yes i havent dig deeper on the algorithm, but based on the mixture distribution, i probably expect the variables to have different variances and covariance so i can make some assumptions on the correlations.
Alright. Have you looked into mixture modeling at all?
I find this paper to be a good introduction:
Article Ten frequently asked questions about latent class analysis
It sounds like you could start by setting up a latent profile analysis, where means, variances, and covariances are are freely estimated within and across profiles.
Thank you for the link. Yes, I have looked into mixture model, and I am going to study further on Latent Tree Analysis since i'm doing multi clustering. Still finding out ways on modifying the algorithm to suit my area of study.
- Badges
- Science topic