Applications to Corporate Default Prediction and Consumer Credit
This chapter illustrates the application of analytical predictive and descriptive techniques for credit risk assessment. To this end, two case applications are presented using data sets involving corporate defaults and credit card loans. The first part is devoted to the prediction of corporate defaults. A data set of 13,414 European small and medium-sized manufacturing enterprises (SMEs) from six countries is considered during the period 2009–2011. The information available for the firms involves their financial characteristics. Corporate default prediction models are constructed with statistical, machine learning, and multicriteria decision making techniques. The analysis of the results covers both the predictive performance of the models, as well as the insights that they provide regarding the factors that affect the default risk for European SMEs. In the second part, a descriptive multivariate clustering approach is employed to obtain analyze credit card loan applications. A publicly available data set of 30,000 cases is analyzed with the k-medoids algorithm to identify clusters of borrowers having similar characteristics. The results are discussed in terms of the common features of the clusters and their level of credit risk.
- Akkoç, S. (2012). An empirical comparison of conventional techniques, neural networks and the three stage hybrid Adaptive Neuro Fuzzy Inference System (ANFIS) model for credit scoring analysis: The case of Turkish credit card data. European Journal of Operational Research, 222(1), 168–178.CrossRefGoogle Scholar