Credit Rating Prediction by Using Machine Learning Methods

Abstract

In this thesis, we discuss the credit ratings of companies. Our purpose is to make a credit rating prediction rule that gives each company a credit which is as correct as possible to the actual rank. We describe three representative machine learning algorithms, which are ordinal logistic regression, neural networks and support vector machine. In addition, we try to analyze their performance and correctness and compare them to determine which method is the most efficient in machine learning to decide ratings. We deal with two different data sets of experiments which consist of true credit rating of companies in 2009 and 2013 and financial information in the previous year.