Yardstick metrics in `fit_resample` always OOS?

[kydare]

Good day,

I'd like to ask whether fit_resample returns out-of-sample results for all the possible metrics. In particular, I'm using yardstick::rsq to evaluate a GLM and get rather good results. Unfortunately, I didn't find a way to also extract training metrics to check myself.

Thank you very much in advance,
Kydare.

[Max]

Unless you use the apparent() function as your resampling object, the results are all out of sample.

You could do it with an extract argument to the control function. You would make a function that predicts the training set and computes the metrics.