Use the churn data set, given at the book series website, for the following exercises. Use the Churn_Training_File. Filter out all variables except the following: VMail Plan, Intl Plan, CustServ Calls, and Churn. Set CustServ Calls to be ordinal. Allow the three predictors to be in either antecedent or consequent, but do not allow Churn to be in the antecedent.

Set the minimum antecedent support to 1%, the minimum rule confidence to 5%, and the maximum number of antecedents to 1.

a. Generate rules using confidence difference as your evaluation measure with evaluation measure lower bound = 40. Explain what this evaluation measure means.

b. For the rules that are generated, use hand calculations to compute the reported evaluation measure, and show that the evaluation measure lower bound has been met.

c. Generate rules using confidence difference as your evaluation measure with evaluation measure lower bound = 30.

d. Select the rule with the highest deployability. Explain why the deployability of this rule is greater than the rule we found in Question 1a.