Fill in the blank: is defined as data is held aside during the modeling process and used only to evaluate after the modeling is complete. Big data Training data Unstructured data Test data QUESTION 6 Which of the following techniques can be used to avold overfitting when using a gradient boosting model to predict customers' purchase likelihood? L. fegularization L2 regularization Early stopping All of the above QUESTION 7 Which of the following statement is consistent with the intuition of collaborative filtering: Person A purchases apple and banana. If person B purchases apple, banana, and coconut, then person A should be more likely to purchase coconut. Person A purchases apple and banana. If person B follows person A on Instagram, B should be more likely to purchase apple. If person A purchases a lemon, A is more likely to buy a lemon flavored soda. H person A purchases a lemon, then person B will also buy a lemon