[10 pts.] Consider a training set that contains 100 positive examples and 400 negative examples. For each of these candidate rules: .Ri-positive class, covers 4 positive and 1 negative examples; - R2-positive class, covers 30 positive and 10 negative examples; R3-> positive class, covers 100 positive and 90 negative examples; determine the best rule according to (show your work): a.) rule accuracy b.) "improved" likelihood ratio