Explaining Preferences with Shapley Values Robert Hu