I would like an answer to part 2 of this question.

In this question you will analyze one possible approach to the question of how to split a dataset S into training and test sets, Sun" and 5\"\". As we have already discussed, overly small test. sets lead to unreliable loss estimates, whereas overly large test sets leave too little data for training, thus producing poor prediction models. The optimal trade-off depends on the data and the prediction model. So can we let. the data speak for itself? we will give it a try. 1. To warm 11p: assume that you have a xed split of S into Strain and 3\"\