Evaluation and Comparison of Classification Model Performance in Predicting Corporate Credit Ratings Using Artificial Intelligence: A Case Study of the Tehran Stock Exchange

This article examines and evaluates the performance of four classification models in predicting corporate credit ratings. The models under study include Support Vector Machine (SVM), Artificial Neural Network (Neural Network), k-Nearest Neighbors (KNN), and Decision Tree. The data used includes features such as exports, company age, production volume, external auditing, foreign ownership, ownership type, and company size, all extracted from the financial statements of companies listed on the Tehran Stock Exchange. The data was divided into training and testing sets, standardized, and then used for training and evaluating the models. The performance of the models was assessed based on accuracy, precision, recall, ROC AUC score, and confusion matrix. The results indicate that the Decision Tree model, with an accuracy of ۱.۰۰۰ and an ROC AUC score of ۱.۰۰۰, exhibited the best performance in predicting corporate credit ratings. The SVM and Neural Network models demonstrated very good performance with an accuracy of ۰.۹۹۵ and an ROC AUC score of ۰.۹۹۹. The KNN model showed acceptable performance with an accuracy of ۰.۹۹۰ and an ROC AUC score of ۰.۹۹۳. This study demonstrates that classification models can effectively aid in predicting corporate credit ratings, with the Decision Tree model being identified as the best option in this context.

کلیدواژه ها:

Corporate Credit Ratings ، Support Vector Machine ، Neural Network ، K-Nearest Neighbors ، decision tree ، Performance Evaluation ، Tehran Stock Exchange