(Choose 1 answer)
True or false? Assume we are using word n-grams as features to perform sentiment classification. Then,higher values of n will usually be less prone to overfitting (i.e., for higher values of n, the difference between training and testing accuracies will be smaller).
A. True
B. False
18
