Outliers in GMM PreviousNext
Mplus Discussion > Latent Variable Mixture Modeling >
Message/Author
 milan lee posted on Monday, December 23, 2013 - 7:19 am
Hi,
I am confused at the issue of outliers in GMM. How should we treat outliers? Should they be deleted prior to modelings?I doubt so but have no idea what other ways to take care of outliers.Thank you very much!
 Bengt O. Muthen posted on Monday, December 23, 2013 - 5:00 pm
Page 768 of the UG discusses outlier options.
 Asgeir Røyrhus Olseth posted on Thursday, February 04, 2016 - 12:30 am
Dear drs. Muthén,
I´ve read the section of the UG pertaining to outliers, but there is one basic thing I hope you can help clarify for me. We´re analyzing data about externalizing behavior in a sample of 4th and 5th graders who we´ve followed through 7th grade. We expect relatively few individuals in some of the classes, and see that a cases “outlier status” (based on Influence) is partly decided by the number of classes.

So my question is therefore when in the process outliers should be identified and removed? Is it before, during or after the class enumeration process?

Best regards,
Asgeir R Olseth
 Bengt O. Muthen posted on Friday, February 05, 2016 - 6:08 pm
I think with mixtures it needs to be during the class enumeration process. Outliers can be forming a class after all.
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: