Non-negative matrix factorization with sparseness constraints for credit risk assessment

As the most important tasks of a bank, assessment of credit card users is aimed to keep the risk of a credit loss low and to minimize costs of failure over risk groups. Credit risk assessment is an essential problem in finance. However, accessing credit risk is very difficult because many factors may contribute to the risk and their relationship is complicated to capture. Recent years have witnessed a growing trend in applying machine learning methods, such as SVM classifier, for credit risk analysis. SVM is a strong classifier that is effective in capturing nonlinear relationship in the data. However, high dimensional training data not only results in time-consuming computation but also affects the performance of the classifier. In this paper, we will adopt sparse non-negative matrix factorization to transform the data into lower dimensional space that will contribute to good performance in the credit risk classification. We test our method in a real-world credit risk prediction task, and our empirical results demonstrate the advantage of our method by comparing with other state of art methods.