Reassessment and monitoring of loan applications with machine learning

Official URL: http://risc01.sabanciuniv.edu/record=b1655740 (Table of Contents)

Credit scoring is one of the most important dimensions of the decision-making process for the loan institutions. It focuses on quantifying the credit worthiness of applicants. In the first part of the thesis, we aim to investigate the role of machine learning for the reassessment of the applicants. In a sense, we propose a complementary screening step to an existing scoring system. We use a real data set from one of the prominent loan companies in Turkey. The information provided by the applicants form the variables in our analysis. The company’s experts have already labeled the clients as bad and good according to their ongoing payments. Using this labeled data set, we execute several classification algorithms to determine the important variables. As the data set consists of applicants who have passed who have passed the initial scoring system, most of the clients are marked as good. To deal with this imbalanced nature of the problem, we employ a set of different approaches to improve the performance of predicting the applicants who are likely to default. In the second part of this thesis, we aim to predict the payment behaviors of the clients based on their static and dynamic information. Behavioral scoring focuses on the financial risk after the loan decision. It corresponds to monitoring the payment behaviors of the existing clients and taking the related actions when necessary. We consider the payment behaviors as well as the general demographic and financial information of the clients. We employ several classification methods. Furthermore, we also analyze the effect of the length of the event history and the staying power of the prediction models.