Question 6 (20 marks) 1000 customers are selected from a bank's data base for credit card promotion. Two data science models are developed to predict whether they will be interested in their new high- return saving plan. The confusion matrices for their classification performance were shown below: Model A Model B Predicted Predicted + + 100 70 + 130 40 Actual Actual 30 800 60 770 a) Calculate all important metrics for the confusion matrix of model A and B. (5 marks) b) Discuss any two most relevant metrics that should be optimized in the above situation. (4 marks) c) Compare the performance of the models and choose the best one for deployment. (5 marks) d) If the training and testing datasets for the model you chose do not agree with each other, suggest the possible causes. (4 marks) e) Estimate the percentage population that are interested in the new product. (2 marks)