The Winnow algorithm is used for this task, models for which are often tested using the Iris dataset. (-5[2])This task is the alphabetically first of the two most common applications of data structures constructed by the C4.5 and ID3 algorithms. One model for this task may seek to minimize the hinge loss on linear data or utilize the kernel trick on nonlinear data; that model determines the maximum hard or soft margin using (*) support vectors. Due to their namesake convergence theorem, perceptron models are guaranteed to locate separating hyperplanes in this task for linearly separable datasets. Generative models for this task include naive Bayes, in contrast with discriminative models like logistic regression that are used for its binary form. (-5[1])Spam filtering is a classic example of, for 10 points, what task in machine learning that seeks to categorize data? (10[1]0[1])■END■