Comparing Model Performances Using JMP® Pro 10: Predicting Whether a Customer Would Subscribe to a Term Deposit When Contacted for a Bank Direct Marketing Campaign
Dinesh Yadav Gaddam; Krishna Reddy Gade, Students, Oklahoma State University
“Who is the potential customer?” - has always been the million dollar question for a marketing manager. Companies invest considerable amounts of budget in promoting their products but the real challenge has always been to achieve a good Return On Marketing investment (ROMI). The answers to this question has led to the research of an approach which improves the ability of a company to pick a pool of selected customers and direct their promotions to them. This approach is called directed marketing. Through this poster we try to demonstrate the capabilities of Model Comparison algorithm provided by JMP Pro 10 and come up with the best model which would predict whether a customer would subscribe to a term deposit or not when contacted through the directed marketing approach. The data used is real world data obtained from UCI learning repository. The data was collected from a Portuguese marketing campaign relating to bank term deposit subscription. The dataset contains 45211 observations, 16 Input variables and a target variable and contains no missing values. No multi-collinearity was observed among input variables and important variables for model building were identified using the chi-square test.