To contact us with non-public matters, please write to help@r-inla.org or hrue@r-inla.org , otherwise you can use the discussion forum.
If you need help to setup a model or if you are experience some strange behaviour or errors in R-INLA, please contact us at help@r-inla.org
If you want to report an error in the sense that you program crash, do not return any results or give some strange error messages, we need to be able to rerun your model here. In order to do so, we either need the R-code and the data required, or a copy of the internal model-definition-files itself.
R-code and the data can be sent to the addresses above or the discussion-forum, depending on if you want to do this private or public. If you do not want to send private data, then if the problem can be reproduced with simulated/artificial data then you can use these instead.
An alternative which both keeps the data almost secret and avoids sending messy R-code, is to send us the internal model-definition-files. You can create this easily using
inla(....., keep=TRUE, inla.call="")
With these options, the files will be create but the Bayesian inference will not be performed. In the current (seen from within R) directory/folder, look for a new directory/folder named
inla.model
and just send us that directory/folder. If you have one of these from before, then the new models are created as inla.model-1, inla.model-2, etc... So just send us the most recent one. With this option, your data stays almost secret, in the sense that they are there, but in a form that makes it very hard to extract any useful information about unless knowing how the model was defined. However, it contain in most cases what we need to rerun and check what is going on.