Closed Goschjann closed 6 years ago
Done, included binary knn classifier as example for train test error distinction on second slide.
Should we add an extra explanation slide or do you simply describe it in the videos? @berndbischl @christophM
@Goschjann It should be okay without an explanation. I changed k to 3 and added made some stylistic changes. With k=3 it is less confusing to understand why a prediction was a certain way, because there are no ties now.
QDA, train data, fitted modell: estimate for test error vs. estimate for train error