A bank wants to predict the creditworthiness of its customers. Based on the customer records, the credit history, etc., a customer should be classified as creditworthy or unworthy of credit. It is five times more ‘expensive’ for the bank to rate a customer who is unworthy of credit as creditworthy than vice versa. In addition, not all information is available for all customers. For 1,000 representatively selected customers, the creditworthiness is known. For these customers the following data has been collected. (Features for which not all values are known are marked with the addition “incomplete”.)

This is to develop a predictive model that assesses the creditworthiness of future clients. It can be assumed that for these clients as well, there are missing values in the features “Purpose”, “Present employment since”, “Job” and “Foreign worker”