Key Determinants of Credit Default in Personal Loans: A Comparative Analysis of Logistic Regression and Random Forest

DOI

10.2991/978-94-6239-830-6_2How to use a DOI?

Keywords

Credit default prediction; Logistic regression; Random forest; Feature importance; Personal loans

Abstract

Accurate credit default prediction is a core aspect of consumer credit risk management. This study based on a publicly available personal loan dataset containing 32,581 records, systematically compares the performance of logistic regression and random forest in default prediction through data preprocessing, exploratory data analysis, feature selection, and classification modeling. It focuses on examining the relative contributions of three types of variables—borrower basic information, loan application information, and credit history information, to default risk. The study finds that indicators reflecting repayment burden (loan amount as a percentage of income) are the strongest determinant of default prediction, followed by loan pricing (interest rate, credit rating). Borrower income and housing situation are the second most important dimensions, while the incremental predictive effect of credit history length is limited. Model comparisons show that random forest outperforms logistic regression in accuracy, AUC, and F1 score, indicating that default risk is significantly affected by nonlinearity and interaction effects; however, logistic regression still holds significant value in practice due to its interpretable coefficients. The study strengthens the internal validity of the conclusions through multiple robustness tests, including threshold adjustment, multicollinearity diagnosis, permutation importance, and SHAP interpretation. This research provides empirical evidence for financial institutions to optimize pre-loan risk assessment and variable selection.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Yi Shen
PY  - 2026
DA  - 2026/11/14
TI  - Key Determinants of Credit Default in Personal Loans: A Comparative Analysis of Logistic Regression and Random Forest
BT  - Proceedings of the 6th International Conference on Internet Finance and Digital Economy (ICIFDE 2026)
PB  - Atlantis Press
SP  - 4
EP  - 26
SN  - 2352-5428
UR  - https://doi.org/10.2991/978-94-6239-830-6_2
DO  - 10.2991/978-94-6239-830-6_2
ID  - Shen2026
ER  -