An Interpretable Machine Learning Framework for Loan Default Risk Prediction in Financial Systems
International Journal of Creative and Open Research in Engineering and Management · 08.05.2026 · 02 (05) · 1-9
DOI: 10.55041/ijcope.v2i5.265 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Creative and Open Research in Engineering and Management, online erschienen am 08.05.2026. Verfasst von Ramya P Ramya P und 3 weiteren Autor:innen.
Abstract (Originalsprache)
Reliable identification of loan default risk constitutes a strategic priority for financial institutions seeking to limit credit exposure while sustaining sound lending practices. This paper introduces a structured, machine learning–driven framework built on real‐world financial data to address this challenge. The pipeline encompasses thorough data preparation—including missing‐value removal, binary label encoding, categorical‐to‐numerical conversion, and class‐imbalance correction—to ensure that each model receives consistently formatted, representative input. Three well‐established classifiers—Logistic Regression, Random Forest, and XGBoost—were trained and evaluated across a standard set of performance indicators: overall accuracy, area under the receiver operating characteristic curve (ROC‐AUC), precision, recall, and F1‐score. XGBoost delivered the strongest accuracy by leveraging its gradient‐boosting architecture to capture nonlinear interactions within the data. Complementing the predictive component, an explainability layer based on SHapley Additive exPlanations (SHAP) was integrated to surface the relative contribution of each input variable to individual predictions, thereby converting an otherwise opaque ensemble into a transparent decision‐support tool. The resulting framework strikes a deliberate balance between predictive power and model transparency—a pairing that is particularly valuable in regulated credit‐risk environments where decisions must be both accurate and justifiable. Keywords: Loan Default Prediction, XGBoost, Random Forest, Logistic Regression, SMOTE, SHAP, Credit Risk, Interpretable Machine Learning
Zitieren
APA
P, R. P. R., M, P. D. M. P. D., V, N. G. V. N. G., & N, N. H. N. N. H. (2026). An Interpretable Machine Learning Framework for Loan Default Risk Prediction in Financial Systems. International Journal of Creative and Open Research in Engineering and Management, 02(05), 1-9. https://doi.org/10.55041/ijcope.v2i5.265
Vancouver
P RPR, M PDMPD, V NGVNG, N NHNNH. An Interpretable Machine Learning Framework for Loan Default Risk Prediction in Financial Systems. International Journal of Creative and Open Research in Engineering and Management. 2026;02(05):1-9. doi:10.55041/ijcope.v2i5.265
BibTeX
@article{P2026,
author = {P, Ramya P Ramya and M, Pruthvi D M Pruthvi D and V, Nandini G V Nandini G and N, Nayana H N Nayana H},
title = {An Interpretable Machine Learning Framework for Loan Default Risk Prediction in Financial Systems},
journal = {International Journal of Creative and Open Research in Engineering and Management},
year = {2026},
volume = {02},
number = {05},
pages = {1--9},
doi = {10.55041/ijcope.v2i5.265}
}
EUVIMED