Research on Credit Card Default Prediction for Class-Imbalanced Datasets Based on Machine Learning

Jinyang Liu

2024

Abstract

It’s known that a robust credit relationship is advantageous for both parties involved. However, credit defaults significantly amplify risk for financial institutions. Hence, default rate prediction stands as a crucial objective for lending institutions and a well-functioning predictive model serves as a potent means to strengthen risk control. To this end, this paper constructed multiple machine learning classification models to achieve credit card default prediction. Feature selection, based on the importance of variables in the random forest, was implemented to enhance the model performance. The results shown that, addressing the skewed nature of credit default data, various SMOTE-based resampling methods were employed to improve data distribution and further optimize accuracy. Compared to other models, the random forest model demonstrated superior predictive effectiveness. After correcting the data distribution, there was a significant enhancement in the predictive performance of all models, with K-Means SMOTE showcasing outstanding performance in data correction and model accuracy optimization.

Paper Citation

in Harvard Style

Liu J. (2024). Research on Credit Card Default Prediction for Class-Imbalanced Datasets Based on Machine Learning. In Proceedings of the 1st International Conference on Data Science and Engineering - Volume 1: ICDSE; ISBN 978-989-758-690-3, SciTePress, pages 441-447. DOI: 10.5220/0012818100004547

in Bibtex Style

@conference{icdse24,
author={Jinyang Liu},
title={Research on Credit Card Default Prediction for Class-Imbalanced Datasets Based on Machine Learning},
booktitle={Proceedings of the 1st International Conference on Data Science and Engineering - Volume 1: ICDSE},
year={2024},
pages={441-447},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012818100004547},
isbn={978-989-758-690-3},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Data Science and Engineering - Volume 1: ICDSE
TI - Research on Credit Card Default Prediction for Class-Imbalanced Datasets Based on Machine Learning
SN - 978-989-758-690-3
AU - Liu J.
PY - 2024
SP - 441
EP - 447
DO - 10.5220/0012818100004547
PB - SciTePress