Thanks, following your argument 10 times 10-fold cross-validation is as good as 10 times cross-validation with a 90/10 split. Thus, when repeating CV, using folds is not necessary, right?

But what about the overlap of training/testing data when repeating cross-validation? Doesn't this have some negative implications?

@Sebastian The output neuron is linear, only the hidden neuron sigmoid.

You give examples, but don't explain what the exact difference is. Could you please elaborate? Thank you!

I have another question, just to be clear: The second picture alone could be an example for spatial dependency as well as(!) spatial heterogeneity, right? As long as we do not know the process itself we cannot tell from the data, right? Thank you very much!

Connections can be non-unique (see ieeexplore.ieee.org/document/714176). Hence, is it okay to simply count the connections? Maybe we should distinguish between parameter and hyperparameter?