A classifier was built to distinguish between 3 types of news articles: S, H, and P. After training, it was run over 15 testing news articles and here are the results: Article # Actual type Predicted type Article # Actual type Predicted type 1 S P 9 P P 2 P H 10 H P 3 H P 11 P P 4 P S 12 H S 5 S S 13 P H 6 H H 14 S S 7 S S 15 H P 8 H H a) Compute the confusion matrix for the classifier. b) Compute Macro-F1 for the classifier. Show your steps.