Predict first: As k increases from 3-fold toward leave-one-out, what happens to the number of training examples used in each fit, and what happens to the number of models we must train?
Current fold
Used for fitting in this foldHeld out in this foldModel fitted without held-out points
Fold errors
| Fold | # train | # held out | MSE |
|---|