Please answer the questions Q1, Q2, Q3, and Q4 based on the information provided below: The XYZ company, an online drugstore, wants to improve the efficiency and effectiveness of its direct marketing efforts (e.g., sending customized coupons, catalogs etc.). 10,000 customers. The first step in this process is to identify the "active" customers. To that end, they conduct a series of logit analyses using a dataset of randomly selected sample of The variables used in the analyses are: 1. Choice (i.e., Response): 1, if customer is assumed to be active, 0 = otherwise 2. Tsince_lastorder: Number of months since the last order of the customer 3. avgorder: Average order value for the customer 4. ordfreq: Number of orders per month 5. esent: Number of emails sent 6. refill: 1 if customer subscribed for automatic refill 7. doorstep: 1 if customer subscribed for doorstep delivery 8. paperless: 1 if customer subscribed for paperless communication (only online) Calibration data Calibration data contains 8000 rows and 8 columns. It contains 7 predictors, to which we added an intercept. Click Save and Submit to save and submit. Click Save All Answers to save all answers. Save All Answers Save and SubmitModel statistics Model parameters reported here have been estimated on the entire calibration data. Parameters Standard deviations P-values Intercept 7.935 0.4567 0.0000 Tsince lastorder 1.066 0.0888 0.0000 avgorder -0.001 0.0010 0.1371 ordfreq -0.641 0.6388 0.3158 esent 0.274 0.0088 0.0000 refill -0.026 0.4957 0.9588 doorstep -0.447 0.7404 0.5464 paperless -1.048 0.2028 0.0000 Model statistics. For indentification purpose, parameters for alternative 0 have been fixed to 0. They are not reported here. P-values = probabilities that parameter estimates are different from zero only by chance. What can you say about the predictive power of the logit model? Click Save and Submit to save and submit. Click Save All Answers to save all answers. Save All Answers Save and SuWhat can you say about the predictive power of the logit model? For the toolbar, press ALT+F10 (PC) or ALT+FN+F10 (Mac). BIUS Paragraph V Arial 14px EV A v ... P O WORDS POWERED BY TINY QUESTION 2 10 points Save Answer Should we use the results of this logit model in our marketing efforts? Why/Why not? (HINT: Your response should be based on the criteria presented in Week 9 software tutorial video) For the toolbar, press ALT+F10 (PC) or ALT+FN+F10 (Mac). BIUS Paragraph Arial 14px Ev A v & T & J Q Q - E ... O WORDS POWERED BY TINY P 10 points Save Answer QUESTION 3 Interpret the "Coefficient Estimates" as if you are reporting it to your manager who knows ABSOLUTELY nothing about statistical significance and only interested in a to- do list! Please DO NOT WRITE A LONG PARAGRAPH, WRITE YOUR ANSWERS IN BULLET POINTS Click Save and Submit to save and submit. Click Save All Answers to save all answers. Save All Answers Save and SubmitQUESTION 3 10 points Save Answer Interpret the "Coefficient Estimates" as if you are reporting it to your manager who knows ABSOLUTELY nothing about statistical significance and only interested in a to- do list! Please DO NOT WRITE A LONG PARAGRAPH, WRITE YOUR ANSWERS IN BULLET POINTS For the toolbar, press ALT+F10 (PC) or ALT+FN+F10 (Mac). BIUS Paragraph V Arial 14px Ev EV AV AV IX ... P O WORDS POWERED BY TINY QUESTION 4 10 points Save Answer Based on the confusion matrix below, please: A) Calculate the hit rate for the whole sample B) Calculate the hit rate for the group of customers who are assumed to be "active" Please EXPLAIN and SHOW YOUR WORK Prodisted n Prodictod 1 Total Click Save and Submit to save and submit. Click Save All Answers to save all answers. Save All Answers Save and Submit\f