Corporate Credit Rating Prediction Using Explainable AI

Abstract

­­­­­­­­­­­Credit ratings are independent opinions expressed by rating agencies on a company's risk profile and future financial commitments. Artificial intelligence (AI) has become increasingly popular for credit assessment, with neural networks and support vector machines offering superior accuracy. This paper analyzes datasets from seven US-based industrial sectors and uses a hybrid ensemble learning model using six machine learning models, including Random Forest, Naïve Bayes, k-Nearest Neighbors, Decision Tree, Support Vector Machine, and Logistic Regression to distinguish between investment and non-investment grades. The hybrid model works best for the DURABLES sector, followed by TELECOM and HEALTH sectors. Explainable AI (XAI) tools like LIME and SHAP explain the prediction outcome of investment-grade and non-investment-grade credit ratings classification. The paper also compares the performance of the hybrid model with eight other related datasets for assessing credit ratings.