Credit Risk Prediction Using Machine Learning and Deep Learning - A Study on Credit Card Customers

Aug 3, 2026

Upload by Noah

Credit card spending growth increases consumer credit exposure and forces banks and financial institutions to manage associated credit risks. Accurate classification of credit card customers as “good” or “bad” supports better underwriting and reduces capital loss. The study evaluates multiple machine-learning approaches—including neural networks, logistic regression, AdaBoost, XGBoost, and LightGBM—using performance metrics such as accuracy, precision, recall, F1 score, ROC, and MCC. Results show XGBoost delivers the strongest performance and improves default prediction for informed lending decisions.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages33

Reading Time1 h 23 min

AccessFree to download

Introduction

Credit risk and credit scoring

Customer segmentation and study goals

Methods

Machine-learning and deep-learning models

Evaluation metrics

Results and Discussion

Model comparison and best-performing approach

Conclusion

Implications for lending decisions

Article Credit Risk Prediction Using Machine Learning and Deep Learning: A Study on Credit Card Customers Victor Chang 1, *, Sharuga Sivakulasingam 1, Hai Wang 2, Siu Tung Wong 3, Meghana Ashok Ganatra 1 and Jiabin Luo 1 1 Department of Operations and Information Management, Aston Business School, Aston University, Birmingham B4 7ET, UK; [sharuga1225@gmail.com](sharuga1225@gmail.com) (S.S.); [meghana.ganatra@gmail.com](meghana.ganatra@gmail.com) (M.A.G.); [j.luo2@aston.ac.uk](j.luo2@aston.ac.uk) (J.L.) 2 School of Computer Science and Digital Technologies, Aston University, Birmingham B4 7ET, UK; [h.wang10@aston.ac.uk](h.wang10@aston.ac.uk) 3 Institute of Finance and Technology, University College London, London WC1E 6BT, UK; [tommywong962@gmail.com](tommywong962@gmail.com) * [Correspondence: v.chang1@aston.ac.uk or victorchang.research@gmail.com](Correspondence: v.chang1@aston.ac.uk or victorchang.research@gmail.com) Citation: Chang, Victor, Sharuga Sivakulasingam, Hai Wang, Siu Tung Wong, Meghana Ashok Ganatra, and Jiabin Luo. 2024. Credit Risk Prediction Using Machine Learning and Deep Learning: A Study on Credit Card Customers. Risks 12: 174 . [https://doi.org/10.3390/](https://doi.org/10.3390/)[ ](https://doi.org/10.3390/)risks12110174 Academic Editor: Mogens Steffensen Received: 3 September 2024 Revised: 17 October 2024 Accepted: 30 October 2024 Published: 4 November 2024 Copyright: © 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://](https://)[ ](https://)[creativecommons.org/licenses/by/](creativecommons.org/licenses/by/)[ ](creativecommons.org/licenses/by/)[4.0/](4.0/)) . Abstract: The increasing population and emerging business opportunities have led to a rise in consumer spending. Consequently, global credit card companies, including banks and financial institutions, face the challenge of managing the associated credit risks. It is crucial for these institutions to accurately classify credit card customers as “good” or “bad” to minimize capital loss. This research investigates the approaches for predicting the default status of credit card customer via the application of various machine-learning models, including neural networks, logistic regression, AdaBoost, XGBoost, and LightGBM. Performance metrics such as accuracy, precision, recall, F1 score, ROC, and MCC for all these models are employed to compare the efficiency of the algorithms. The results indicate that XGBoost outperforms other models, achieving an accuracy of 99.4% . The outcomes from this study suggest that effective credit risk analysis would aid in informed lending decisions, and the application of machine-learning and deep-learning algorithms has significantly improved predictive accuracy in this domain. Keywords: credit risk prediction; credit risk; classification; machine learning 1. Introduction Credit cards offer an easy method of borrowing money to pay for a range of goods and services in the modern era. Credit cards function as a replacement for cash and debit cards and are also widely used in daily shopping. Moreover, credit cards have become indispensable to the contemporary economy for many individuals. From the view of financial institutions and banks, they must decide whether to approve credit card applications when a customer submits the application. Different factors are considered by creditors when determining whether the customer is good or bad in terms of risk andrepayment. According to practical scenarios, our approach categorizes a “bad customer”as one whose credit has been past due for 60 days or more, whereas a “good customer” isone whose debt has been past due for less than 60 days. Credit scores are a general risk governance strategy in banks and other credit institutions, which utilize the personal data of credit card customers to signify

FAQs

How does the study define “good” and “bad” credit card customers?

A “bad customer” is defined as having credit past due for 60 days or more, while a “good customer” has debt past due for less than 60 days.

Which models are evaluated for credit risk prediction?

The study applies neural networks, logistic regression, AdaBoost, XGBoost, and LightGBM to predict default status.

What evaluation metrics are used to compare the models?

Models are compared using accuracy, precision, recall, F1 score, ROC, and MCC.

Aug 3, 2026

Upload by Noah