You are asked to run a Bayesian classifier on a learning task with two classes C1 and C2. You are given the posterior probabilities for the example X under analysis: P(C1|X) = 0.6, P(C2|X) = 0.4. You are also giving a loss matrix L(Z1, Z2) where Z1 is the true class and Z2 is your prediction. These are the values for the loss matrix: L(1,1) = 0; L(2,2) = 0, L(1,2) = 3; L(2,1) = 5. What is the EPE (expected prediction error) when the predicted class is C1, C2, and what is the final predicted class?