Refer to Exercise 11 for a description of the data set. Partition the data into 60% training and 40% validation data. For Analytic Solver, use 12345 as the random seed and create 10 weak learners. For R, use one as the random seed and create 100 weak learners. 

a. Create a bagging ensemble classification tree model to determine whether a customer will travel within the next year. What are the overall accuracy rate, sensitivity, and specificity of the model on the validation data? What is the AUC value of the model? 

b. Compare the performance of the bagging ensemble model to that of the single-tree model created in Exercise 11 (for Analytic Solver) or Exercise 12 (for R). Which model shows more robust performance? Explain. 

c. Score the two new customers in the Travel_Plan_Score worksheet using the bagging ensemble classification tree model. What is the probability of the first customer having plans to travel within the next year? What is the probability for the second customer?

Data from Exercises 11 

Jerry Stevenson is the manager of a travel agency. He wants to build a model that can predict whether or not a customer will travel within the next year. He has compiled a data set that contains the following variables: whether the individual has a college degree (College), whether the individual has credit card debt (CreditCard), annual household spending on food (FoodSpend), annual income (Income), and whether the customer has plans to travel within the next year (TravelPlan, 1 = has travel plans, 0 = does not have travel plans). A portion of the Travel_Plan_Data worksheet is shown in the accompanying table. Create a classification tree model for predicting whether or not the customer will travel within the next year (TravelPlan). Select the best-pruned tree for scoring and display the full-grown, best-pruned, and minimum error trees.

Transcribed Image Text:

College Credit Card No No Yes No ⠀ B No No FoodSpend 2905.97 4677.03 ⠀ 1892.37 Income 65982 57274 ⠀ 77626 TravelPlan 1 1 B 0