This homework assignment will use the wbca dataset of the faraway package. 1. First recode the response variable variable Class=1-Class so that a 1 refers to a malignant tumor. Then fit a logistic regression model using all of the predictors. a. Compute the measure of predictive power which is the correlation between y and TT . b. Compute the measure of predictive power LS-Lo 2. Using a cutoff of To = .5 produce the confusion table and report the sensitivity, specificity, and correct classification rate. 3. Repeat (2) but using the model resulting from backward elimination with the AIC method. Describe any changes to sensitivity, specificity and correct agreement rate. 4. Using the model of (3) produce the confusion table and sensitivity, speccificity and correct classification rate using both the probit and cloglog links. 5. Which link function works best for this classification problem? Explain