glme can't deal with NaNs?

10 views (last 30 days)
Ane
Ane on 19 Aug 2016
Answered: Qu Cao on 24 Aug 2016
Hi everyone,
I've fitted a lme model with my dataset of 43 features x 208000 samples. There are many NaN values in the dataset, but lme function has dealt with them and has worked succesfully. Now, I'm trying to create another model with glme function, with the same dataset but for a binary response variable. Matlab complains, saying the following:
"Error using classreg.regr.lmeutils.StandardLinearLikeMixedModel/validateX (line 288) X must have rank 25."
The response variable contains only 0 and 1 values, so I really don't see the problem. glme can't deal with NaNs? What can I do?
Thanks!

Answers (1)

Qu Cao
Qu Cao on 24 Aug 2016
According to the error message, it seems that the error is not associated with NaN but the rank of the matrix X.
You can check the linked file of 'line 288' to find the requirements of the input matrix X, which are given in function 'validateX' (line 277 - 290).

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!