What formula should I use in linear mixed effects model fitting?

Hello everyone!

In an experiment, 10 raters rated 15 noises (every rater rated the same noises).

I wanted to create a linear mixed model with 3 predictors, but I am not sure which formula to use in the fitlme function.

So my different noises are actually the grouping variable (1 to 15) and each noise contains 10 ratings (10 different persons).

So firstly I created this table below:

lme_tbl = table(lme_inputs(:,1),lme_inputs(:,2),lme_inputs(:,3),lme_inputs(:,4),lme_inputs(:,5),...

'VariableNames',{'Ratings','X1','X2','X3','NoiseGroup'});

You can see that the X variables are repeated for every group.

lme_IA = fitlme(lme_tbl,'Ratings~ ? ')

I am not sure what formula I should use, since the information on fitlme online doesnt make it perfectly clear for me..

The X variables are not related in any way with the grouping variable, they represent physical qualities..

Thank you in advance for your help!

Greetings

Answer 1

[Sulaymon Eshkabilov]

In this case, you should choose the formulation that would give a most accurate prediction. Either you go with your common understanding of your data or trial & error approach.

If fitlme fails, you can test non-linear fit models using fitnlm() fo test various nonlinear formualtions again using your knowledge or just trial & error approach.