Data Mining Week 8 Instructions Address the following questions in complete sentences and submit the answers in a Word document. Problem 1: You are an operations analyst working for a

major movie theater chain. Currently, the company requires all employees in the concession stand to use a technique known as suggestive selling (e.g., "Would you like Red Vines with your order today?"). Employees are not given any guidelines as to what suggestion to make, so they typically pick their favorite food or candy. The company has asked you to come up with a system to find items that a given customer is likely to buy in order to improve the effectiveness of the suggestive selling process and increase profits. Using market basket analysis, you have discovered several associations that you believe will lead to improved suggestive selling accuracy. You are considering two main options for deployment: 1. Work with the point-of-sale software vendor to implement the suggestive selling algorithm into the software used by cashiers. As cashiers enter the customer's order, the algorithm will determine what additional item the customer is likely to buy, based on the items currently ordered. The item will be displayed to the cashier who can then choose to add the item to the current order or dismiss the prompt, depending on the customer's response. Naturally, the software vendor is demanding a hefty fee and promising a 6-12-month timeline for delivery of the software update. Changes to the algorithm down the road will require an additional fee and waiting period. 2. Choose a small number (e.g., 5-10) of the most effective rules and incorporate them into the training protocol. One of these rules would be a catch-all (i.e., "If no other rules apply, suggest Red Vines"). You estimate that it will take 2-4 weeks to decide on rules, develop training collateral (posters for the employee breakroom, reminder cards to stick on each cash register, etc.), and deliver training. Because this is a high turnover industry and training is constantly being delivered, the additional costs of delivering this training are negligible. In 100-250 words, describe what sort of analysis you would need to do to make an informed decision about what plan to implement. What data would you need to acquire? How would you acquire that data? How would you measure the success of the deployment? How would you determine when the model needs to be updated? Problem 2: You have decided to implement one of the plans from the previous exercise. The company has asked you to write a 250-word e-mail message that will be sent out to all managers in the company announcing the upcoming change. While managers understand movie theater operations, they do not have a background in data mining. Your e-mail should clearly explain all of the following in layman's terms. You may make up figures related to costs, revenue increases, etc., as needed, to support your e-mail communication. 1. Type of model being implemented 2. Purpose for implementing the model 3. Anticipated results 4. Next steps and timeline for implementation Problem 3: Read the case study "Championing of an LTV Model at LTC," located in topic Resources, and answer the following questions in complete sentences. 1. What was the business problem the authors were trying to solve? 2. What type of modeling activities did the authors use? (description, prediction, classification, etc.) 3. How did the authors evaluate their model? 4. What stages of CRISP-DM are not represented in this case study? 5. Based on this article, why do you think it is important that implementers of data mining models possess strong interpersonal skills? 6. Consider the "Data Science Code of Professional Conduct" and evaluate two ethical issues related to data mining and the responsible stewardship of personal information. https://www.researchgate.net/publication/220520009 Championing of_an_LTV_model_ at LTC