Which of the following are accurate descriptions of standard best practices in Machine Learning (as discussed in STA 301)? (1) Data splitting: use the out-of-sample testing data set to fit models, then use the in-sample training data set to evaluate the models' predictive performance. (2) Overfitting: the primary goal is to build a predictive model that memorizes the pattern of random noise in a sample data set. (3) Feature engineering: building large complex models that include as many plausibly relevant variables as possible, as well as their interactions. O 1 and 3 1 only 2 only All of the above (1, 2, and 3) 3 only None of the above 1 and 2 2 and 3