Thanks a lot for your help, Martin. The model you posted computes and provides pretty satisfactory results too.
I have two more questions, related to this model, that maybe you might help me with.
First, I was trying to write this model in mathematical form. Following this section of the radon notebook:
I came up with the following:
Do you think this mathematical representation is correct for the model you sent? What would the matrix P look like?
Also I’m trying to a obtain graphviz representation of this model, but for some reason when I try to use graphviz with covarying models, through the command
pm.model_to_graphviz(covarying_intercept_and_temp)
I get the following error:
AttributeError: 'float' object has no attribute 'name'
It’s a pretty vague error and I can’t seem to debug it. Graphviz works fine with every other model, but as soon as I introduce covarying terms this happens. Might be a long shot but maybe someone here encountered a similar issue at some point.
Thanks a lot for your help again!