Credit companies incur great risk when providing credit to individuals without the certainty that the individual will pay back the amount borrowed. To reduce this risk and ensure that creditworthy customers are given appropriate amounts of credit, such companies use a customer’s previous borrowing information to help predict their future credit behavior in terms of their likelihood of paying back the loan.
In this project, I investigate how a customer’s borrowing and repaying habits interact with their sociological categories to enable an accurate determination of their chance of default in the immediate future (the following month). I also seek to understand this bank's lending characteristics based on the different categories of customers and how these categories performed with regard to paying back their loans. Finally, I hope to find the most accurate classification algorithm to inform the company’s decision to extend credit to customers so as to minimize the risk of default.