auss'rlou a "porn: Answer QB, QB, and Q10 using the information and the software output below: The ABC Bank. wants to sell a new credit cards to its current customers. They want to improve the e'lciency and effectiveness of its direct marketing efforts since the prot margins of tor these accounts are very low. The rst step in this process is to identify the factors that affect the probability of a customer applying for a new credit card. To that end, they conduct a series of logit analyses. The variables used in the analyses are: choice (m); Has the customer applied for a new credit card (1: if the customer apptied & 0: otherwise) Egg Age of the customer marital: Marital status of the customer (1: if the customer is married at 0 otherwise) balance: Average yearly balance. in U33 M Does this customer have a loan from our bank (1: yes at 0: no) contact work: Did we contact this person about our new credit card while she was at work? (1: yes 8! 0: ho) hum calls: Number of contacts performed during this campaign and for this client lest duration: How long did our last call to this customer last? (in seconds) prey outcome: Outcome of the previous marketing campaign to this customer (1: success & 0: failure) Please answer the questions based on the results of the analyses presented below. Calibration data age marital balance loan contact work num calls last duration prev outcome Average 40.86 1 376.2 0 2.855 249.410 Standard deviation 11.10 0 2 789.5 0 3.209 269.207 0 Median 39.00 460.0 0 2.000 164.000 Minimum 22.00 0 -1 350.0 0 O 1.000 7.000 Maximum 80.00 1 27 446.0 30.000 1 925.000 Alternative 0 Alternative 1 Count 351.0 9.0 Frequency 87.8% 12.3% Out-of-sample data age marital balance loan contact work num calls last duration prev outcome Average 39.6 874 0 2.94 266.3 0 Standard deviation 10.6 1 161 0 0 3.02 263.4 0 Median 37.0 472 2.00 159.0 O Minimum 18.0 0 -281 0 0 1.00 11.0 Maximum 66.0 4 787 15.00 1 148.0 Confusion matrix and hit rate Predicted 0 Predicted 1 Actual 0 98% 2% Actual 1 82% 18% Confusion matrix (%). The global hit rate of the model is 88%. The diagonal elements represent alternative- specific hit rates.Model statistics Parameters Standard deviations P-values Intercept -4.45089108 0.28074179 0.0000 age 0.00673502 0.01384074 0.6265 marital -0.41342256 0.35087356 0.2387 balance -0.00005085 0.00007147 0.4767 loan -1.56297523 0.03852300 0.0000 contact work 2.03400039 0.31913162 0.0000 num calls -0.06318733 0.07906071 0.4242 last duration 0.00201443 0.00047593 0.0000 prev outcome 2.21552847 0.30217781 0.0000Gain chart and lift 100% - 75% - 50% - 25% - 0% 0% 25% 50% 75% 100 Top 5% Top 10% Top 25% Random 5.0% 10.0% 25.0% Truth 40.8% 81.6% 100.0% Model 22.4% 34.7% 69.4% Observed lift 4.49 3.47 2.78 Improvement 48.7% 34.5% 59.2% Out-of-sample predictions Alternative 0 Alternative 1 Count 44.00 3.00 Frequency 93.6% 6.4% Frequency table for predicted choices.Q8. Interpret the statistics for the variables in the calibration data USE A SEPARATE PARAGRAPH FOR EACH VARIABLE. For the toolbar press ALT+F10 (PC) or ALT+FN+F10 (Mac), QUESTION 9 10 points Save Answer XYZ Bank hires you as a consultant and asks you to make them recommendations. A) Please explain whether or not it is a good idea for the XYZ bank to use these datasets in their endeavors to improve the efficiency and effectiveness of its direct marketing efforts. (HINT: Use the criteria presented in the video tutorial) B) Based on the "model statistics" table, what recommendations would you have to the management? (HINT: The question does not ask your interpretations of the table, it asks you to come up with recommendations to your clients, who, presumably, have zero knowledge about statistics or logit models)QUESTION 1|] "palm: avenue!" If we want to reach 50% of the \"active" peopleI what percentage 0! our sample we need to use? (After we rank the people In our sample based en their probability of being active using our model} PLEASE EXPLAIN OUR