Combining Categories in a Classification Problem Below is the Classification Matrix (Confusion Matrix) for a model that predicts a categorical dependent variable that can take three possible values: Low (L), Medium (M), High (H). Actual\\Predicted M W I 13 8 M 7 11 UI N H 0 1 Suppose our original categorization is simplified by absorbing the category Medium into the category High. That is, any element that was originally a Medium is now considered a High. And any predicted classification of Medium would instead be treated as a predicted classification of High. (Note: We are not retraining the model, which might produce an entirely different classifier, but just relabeling the elements of the dataset and the outputs of the original model. This new input-output relationship can be thought of as a new classifier; we just do not have an explicit understanding of its structure.) Answer the following questions about the new Classification Matrix, depicted as Actual\\Predicted I H Hint: The correct approach does not entail simply collapsing together two rows (or two columns) in the original matrix. Instead, go through each cell of the 2-category matrix and identify which cell(s) in the 3-category matrix would feed into that cell. To keep you from going wildly astray, note that the correct value of b is 11. To simplify the thought process and provide a shorthand, you may wish to refer to the elements of the original matrix using the letters below. Actual\\ Predicted M I M H 7