Explainable Machine Learning for Credit Risk Prediction as Support for Risk Management Decision-Making
Explainable Machine Learning untuk Prediksi Risiko Kredit sebagai Pendukung Pengambilan Keputusan Manajemen Risiko
DOI:
https://doi.org/10.56667/jveit.v7i1.2433Keywords:
Risiko Kredit, Explainable Machine Learning, XGBoost, SHAP, Manajemen Risiko.Abstract
Risiko kredit merupakan salah satu risiko utama yang dihadapi lembaga keuangan akibat potensi gagal bayar (default) nasabah. Prediksi risiko kredit yang akurat dan transparan diperlukan untuk mendukung pengambilan keputusan manajemen risiko berbasis data. Penelitian ini bertujuan membangun model prediksi risiko kredit menggunakan pendekatan Explainable Machine Learning pada Taiwan Credit Card Clients Dataset. Tahapan penelitian meliputi data preprocessing, penanganan ketidakseimbangan data menggunakan Synthetic Minority Over-sampling Technique (SMOTE), optimasi hiperparameter menggunakan RandomizedSearchCV dengan 3-fold cross-validation, serta pelatihan model menggunakan algoritma Extreme Gradient Boosting (XGBoost). Evaluasi model menunjukkan nilai Accuracy sebesar 80,65%, Precision 59,39%, Recall 39,56%, F1-Score 47,49%, dan ROC-AUC 75,54%, yang menunjukkan bahwa model memiliki kemampuan klasifikasi dan diskriminasi yang cukup baik dalam membedakan nasabah berisiko default dan tidak default. Interpretasi model menggunakan SHapley Additive exPlanations (SHAP) menunjukkan bahwa PAY_0, LIMIT_BAL, AGE, serta riwayat pembayaran dan jumlah pembayaran nasabah merupakan faktor-faktor yang paling berpengaruh terhadap prediksi risiko kredit. Hasil penelitian ini menunjukkan bahwa kombinasi XGBoost dan SHAP tidak hanya menghasilkan model prediksi dengan performa yang baik, tetapi juga meningkatkan transparansi dalam menjelaskan faktor-faktor yang memengaruhi risiko kredit. Temuan tersebut diharapkan dapat menjadi dasar bagi lembaga keuangan dalam menyusun strategi manajemen risiko kredit yang lebih akurat, objektif, dan berbasis data.
Downloads
References
[1] S. Kaisar and S. Sifat, “Explainable Machine Learning Models for Credit Risk Analysis: A Survey,” 2023, pp. 51–72. doi: 10.1007/978-3-031-36570-6_2.
[2] S. Crone and S. Finlay, “Instance sampling in credit scoring: An empirical study of sample size and balancing,” International Journal of Forecasting, vol. 28, pp. 224–238, Mar. 2012, doi: 10.1016/j.ijforecast.2011.07.006.
[3] A. Wahid and A. Muliawan, “Strategi Retensi Pelanggan Berbasis Historis: Optimalisasi Model Prediksi Churn Menggunakan Machine Learning,” SemanTIK : Teknik Informasi, vol. 11, no. 2, Sep. 2025, doi: 10.55679/semantik.v11i2.237.
[4] M. Pebriadi, T. Fattah, and P. Salman, Interpreting Credit Risk Prediction with XGBoost and SHAP. 2025, p. 582. doi: 10.1109/ISRITI68345.2025.11393290.
[5] D. Kumar, “Explainable Machine Learning Models for Credit Risk Prediction in Retail Lending: A Comparative Study Using SHAP,” SSRN Electronic Journal, Jan. 2025, doi: 10.2139/ssrn.5341125.
[6] B. Hadji Misheva, J. Osterrieder, A. Hirsa, O. Kulkarni, and S. Lin, Explainable AI in Credit Risk Management. 2021. doi: 10.48550/arXiv.2103.00949.
[7] N. Bussmann, P. Giudici, D. Marinelli, and J. Papenbrock, “Explainable Machine Learning in Credit Risk Management,” Computational Economics, vol. 57, Jan. 2021, doi: 10.1007/s10614-020-10042-0.
[8] D. A. Fauziah, A. Muliawan, and M. Dimyati, “IMPLEMENTATION OF MACHINE LEARNING ON EMPLOYEE ATTRITION BASED ON PERFORMANCE PARAMETERS USING PARTICLE SWARM OPTIMIZATION AND ENSEMBLE CLASSIFER METHODS,” Jurnal Teknik Informatika (Jutif), vol. 5, no. 6, Art. no. 6, Dec. 2024, doi: 10.52436/1.jutif.2024.5.6.3442.
[9] L. Shu, C. Su, and J. Shu, A Comparative Study of Explainable Machine Learning Models for Corporate Credit Scoring. 2025, p. 6. doi: 10.1109/ICECET63943.2025.11472005.
[10] D. A. Fauziah, A. Muliawan, I. Sabilirrasyad, and B. W. Maulana, “Smart Campaign for Smart Business Using Machine Learning on Improving Product Marketing Strategy,” Proceeding International Conference on Economics, Business and Information Technology, vol. 6, no. 1, pp. 616–622, Sep. 2025, doi: 10.31967/icebit.v6i1.1681.
[11] S. S and S. M, “Smart and Explainable Credit Card Fraud Detection Using XGBoost and SHAP,” J. IoT Soc. Mob. Anal. Cloud, vol. 7, no. 2, pp. 155–169, Jul. 2025, doi: 10.36548/jismac.2025.2.004.
[12] Shreya and H. Pathak, Explainable Artificial Intelligence Credit Risk Assessment using Machine Learning. 2025. doi: 10.48550/arXiv.2506.19383.
[13] H. Wicaksono, A. Riyanto, R. Darmawan, M. F. Hidayat, and A. Khumaidi, “Explainable Boosting Machine for Transparent Risk Assessment in BAZNAS Microfinance Desa,” ILKOM Jurnal Ilmiah, vol. 17, no. 3, pp. 312–322, Dec. 2025, doi: 10.33096/ilkom.v17i3.3214.312-322.
[14] H. Putra and R. Rumini, “Comparative Study of Logistic Regression, Random Forest, and XGBoost for Bank Loan Approval Classification,” Journal of Applied Informatics and Computing, vol. 9, no. 5, pp. 2822–2835, Oct. 2025, doi: 10.30871/jaic.v9i5.10862.
[15] S. Lin, D. Song, B. Cao, X. Gu, and J. Li, “Credit risk assessment of automobile loans using machine learning-based SHapley Additive exPlanations approach,” Engineering Applications of Artificial Intelligence, vol. 147, p. 110236, May 2025, doi: 10.1016/j.engappai.2025.110236.
[16] L. Lin and Y. Wang, SHAP Stability in Credit Risk Management: A Case Study in Credit Card Default Model. 2025. doi: 10.48550/arXiv.2508.01851.
