warning about overconfident model

[kjappelbaum]

warning about overconfident model

[kjappelbaum]

i think I will just run cross validation and check the scale of the mean mae and the mean variance

[kjappelbaum]

ok, implementing a general cross-validation "hidden" from the users is a bit ugly. But if we have the _train and _predict functions implemented, we can run cross-validation by tampering with the sampled
attribute

[kjappelbaum]

what we still need to think about for the cross validation function is

• do we run always only with min(10, len(samples)) test cases or do we use leave-one-out cross-validation on this set?

[kjappelbaum]

should print now something like