Efficiency Comparison of Multilayer Perceptron and SMO Classifier for Credit Risk Prediction

Abstract: Credit Risk Prediction is an important task in any Banking Industry. Identifying the defaulter before giving loan is a crucial task of the Banker. Classification techniques are used to classify the customer, whether he/she is a defaulter or a genuine customer. Determining the best classifier is a critical assignment for any industrialist. It leads to instil different research works for determining the best classifier for the credit risk prediction. This paper analyzes the efficiency of Multilayer Perceptron Classifier and Sequential Minimal Optimization (SMO) Classifier for the credit risk prediction and compares their efficiency through various measures. The German credit data is taken for credit risk prediction and the classification experiment is done using open source machine learning tool.

Keywords: Credit Risk Prediction, Multilayer Perceptron Classifier, SMO Classifier, Performance Evaluation