In the context of this exercise, why add more dimensions to a predictor?
Adding more dimensions helps increase the accuracy of the predictor.
Each dimension, or feature, that we add to the predictor increases the likelihood of a correct classification.
Take, for example, our dataset of movies, but consider that we only used one feature, the year. Every year multiple movies can be released, but they might have no other similarities otherwise, making it a poor predictor. As we further add other dimensions, say the budget, genre, or director name, we start to get a much more accurate predictor.