CREDIT RISK ASSESSMENT USING DEEP LEARNING TECHNIQUES
CREDIT RISK ASSESSMENT USING DEEP LEARNING TECHNIQUES

Summary/Abstract: Credit risk assessment models are dealing with ranking prospective bank clients as eighter defaulters or not. Traditional predictive techniques, such as decision trees (DT), random forests (RF), support vector machines (SVM), multilayer perceptron (MLP), have already been used to classify customers by risk. In this work we use deep neural networks to solve the problem of credit risk assessment and compare their predictive performance with that of traditional methods. Unfortunately, real data from Romanian banks are currently unavailable, so we conduct our experiment son three publicly available data sets in the UCI Machine Learning Repository. The deep neural networks lead to an average accuracy superior to those obtained with the classical methods mentioned above, for the three databases studied.

• Details
• Contents