Machine Learning in Credit Risk Forecasting —— A Survey on Credit Risk Exposure

Shaoshu Li

Accounting and Finance Research, 2024, vol. 13, issue 2, 107

Abstract: Credit risk is one of the most important elements in risk management area. Traditional regression types of credit risk models are straightforward to implement and model outputs are easy to interpret. However, the model accuracy can always be suboptimal to fit the real credit risk data series. Especially, the model performance even deteriorates under extreme economic scenarios. In contrast, the modern machine learning models can handle different drawbacks of regression types of models. In this paper, we survey the recent literatures on applying the machine learning or deep learning methods in credit risk forecast with special focus on study the superiorities of these techniques. Besides of delivering better prediction accuracies, we uncover other four advantages for machine learning type of default forecast which have been shown in few literatures. We also survey the less studied machine learning or deep learning type of prepayment forecast. By reviewing past literatures from both default and prepayment risk aspects, we can gain comprehensive overview of utilizing machine learning techniques in credit risk forecasting and valuable insights for future risk management research.

Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://www.sciedupress.com/journal/index.php/afr/article/download/25814/15999 (application/pdf)
https://www.sciedupress.com/journal/index.php/afr/article/view/25814 (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:jfr:afr111:v:13:y:2024:i:2:p:107

Access Statistics for this article

More articles in Accounting and Finance Research from Sciedu Press Contact information at EDIRC.
Bibliographic data for series maintained by Sciedu Press ().