Use the accompanying data file to estimate the logistic model for predicting the probability of loan default (Default equals 1 for default, O otherwise). Predictor variables include loan-to-value ratio (LTV in %), FICO credit score (FICO), and customer age. Click here for the Excel Data File a. Use the holdout method, using the cutoff of 0.50, to compute accuracy, sensitivity, and specificity with the first 300 observations for training and the remaining 100 observations for validation. Note: Do not round intermediate calculations and round final answer to 2 decimal places. Performance Measure Accuracy Sensitivity Specificity Value in % b. Compute the corresponding performance measures using the cutoff equal to the proportion of loans with default in the training set. Note: Do not round intermediate calculations and round final answer to 2 decimal places. Performance Measure Accuracy Sensitivity Specificity Value in % c. Which cutoff is preferred in this application, assuming that early identification of potential default is important so that lenders may work with the client to decrease the chance of default? O The cutoff equal to 0.50 because of its higher sensitivity The cutoff equal to 0.50 because of its higher specificity The cutoff equal to the proportion of defaults because of its higher sensitivity O The cutoff equal to the proportion of defaults because of its higher specificity