Page 1 of 1

Fail to store residual for cross-classified model

Posted: Sun Apr 01, 2018 9:17 pm
by vivian1234
Hi, I am running a cross-classified cumulative logit model for my ordinal outcome.
I used the following codes to store the residual:

Code: Select all

estoptions = list(resi.store = TRUE, resi.store.levs = c(2, 3, 4), xc = TRUE, EstM = 1
However, I have this error:
Error: memory exhausted (limit reached?)

In this case, how should I store the residual for residual diagnostic?

Thanks a lot.

Vivian

Re: Fail to store residual for cross-classified model

Posted: Mon Apr 09, 2018 9:02 am
by ChrisCharlton
To ensure the maximum amount of memory is available you could check that you are running a 64-bit version of R, as well as removing unneeded objects from memory (you can list the current objects with the ls() function).

If there still isn't space then you might want to consider whether you need the chains for all of the residuals. You could reduce the information stored by limiting the levels for which you store residual chains or by specifying a thinning value.

Re: Fail to store residual for cross-classified model

Posted: Thu Apr 12, 2018 11:04 am
by vivian1234
Thanks a lot.