Predicting Recovery Rates of Defaulted Credit Card Accounts
MetadataVis full innførsel
Monthly customers are sent to debt collection, where most of them are able to pay the demanded amount and are therefore reinstated as normal customers. However, some do not. These are the ones that constitute the main loss. The problem at hand is therefore to build a model which segments customers into three (or more) groups: the ones who most probably will pay (green), the ones who most probably will not pay (red), and the ones in between (yellow). Obtaining such a model could play an important role in turning yellow customers into green, which yields great profit. Furthermore, a sensitivity analysis of the obtained models will be performed.