V. Problem 5: Classifier Evaluation You have developed two classifiers: C1 and C2 which determine the posterior probability of a student obtaining a pass grade. The results of the 2 classifiers are provided in Table 1. Table 1 True Label C1 C2 1 Pass 0.75 0.80 2 Pass 0.45 0.75 3 Fail 0.20 0.30 4 Pass 0.70 0.65 5 Pass 0.55 0.90 6 Fail 0.65 0.45 7 Fail 0.40 0.65 8 Fail 0.15 0.25 9 Pass 0.95 0.85 10 Fail 0.25 0.60 1. Plot the ROC curves for both the classifiers and determine which classifier is better than the other. (Plot them on the same graph) Show detailed calculations and provide your reasons. 2. A student is speculated to pass if the probability p>=.6. Compute the precision, recall, and F- measure for C1, C2 at this threshold value. Draw the confusion matrix. 3. Suppose you are given a weight matrix as shown in Table 2, use weighted accuracy to determine which classifier is better. Predicted Class Actual Class Yes No Yes 1 2 No 10 1