Predicting Credit Outlook of Banking and Nonbanking Finance Companies: A Comparative Analysis of Machine Learning Models

This article presents a comparative study on forecasting the credit outlook of banking and nonbanking finance companies using five distinct machine learning models: Logistics regression, k-nearest neighbor (kNN), decision tree classifier, support vector machines (SVM), and naïve Bayes classifier. The research evaluates the efficacy of these models in categorizing financial companies based on their credit outlook (negative, neutral, and positive). Our findings reveal that the kNN model with a k value of 7 demonstrated superior performance in terms of test accuracy compared to the other models. Although models such as the kernelized SVM and logistic regression exhibited commendable performance, none could match the high-test accuracy achieved by the kNN model.