How to get R2 and RMSE from my own mixed model analysis

Rizki — Thu, 01 Feb 2018 13:19:13 GMT

Dear Sir/Madam,

I am trying to do the mixed model analysis on meta-analysis study.

But I did not get the R2 and RMSE from the mixed model.

The independent variable is dosage(kg), and some dependent variables (%). Study where classified as a random effect.

I need help for two things:

1. I already had the intercept and slope results, but need the R2 and RMSE results also. Could you please help me how the statement to get it?

2. how to make the statement for graphic estimation based on dependent and the independent variable of the model?

Can anybody help? Please help me as soon as possible.

Many thanks in advance,

Rizki

yanza.draft@gmail.com

Re: How to get R2 and RMSE from my own mixed model analysis

sld — Fri, 02 Feb 2018 00:54:00 GMT

If you are using a procedure like MIXED or GLIMMIX, then a RMSE value is not available because these procedures do not use the least-squares method for estimation.

Edit: The MIXED procedure with the METHOD=TYPE3 option will produce sums of squares and mean squares output based on method of moments estimation. It is not available for models that use a REPEATED statement or have a SUBJECT= effect.

The concept of R2 is complicated for mixed models. This paper and this CrossValidated discussion might be useful to you.

topic Re: How to get R2 and RMSE from my own mixed model analysis in Statistical Procedures

How to get R2 and RMSE from my own mixed model analysis

Re: How to get R2 and RMSE from my own mixed model analysis