Predictive Risk Analytics for Banking Compliance: An Intelligent Framework for Early Detection of Financial and Operational Risks
Journal of Intelligent Decision Making and Information Science · 02.09.2026 · 3 (10s) · 242-271
DOI: 10.17762/jidmis.v3.2711 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Intelligent Decision Making and Information Science, online erschienen am 02.09.2026. Verfasst von Srikanth Girijala und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
MoA dual-stream predictive risk analytics framework is developed and validated for early detection of multidimensional financial deterioration and prospective compliance-enforcement risk in banks. The architecture integrates financial, operational-vulnerability, and historical compliance indicators with temporally protected machine learning, probability calibration, explainability, and structured monitoring, with Random Forest demonstrating the strongest overall validation performance. The calibrated model's performance on the independent sample of financial-risk test data was an ROC-AUC of 0.8741, Average Precision of 0.4509, and a 6.80-fold Average Precision lift over a 6.63% event prevalence. With the operating point selected during validation, the recall was 0.7455 with only 20.28% of the observations being alerted. For compliance risk, the independent test sample yielded an ROC-AUC of 0.7820, Average Precision of 0.1070, and a 10.38-fold Average Precision lift against an event prevalence of 1.03%. At the validation-selected operating threshold, the model captured 44.07% of qualifying enforcement events occurring within the subsequent 12 months while flagging only 4.70% of the test observations. The risk categories ordered by their risk levels revealed clear gradients in event rates, and the SHAP analysis revealed that the top signals for predicting the events were liquidity, profitability, capitalization, asset quality, provisioning, and operational vulnerability. Distinct from conventional failure-oriented or single-domain banking risk models, the proposed dual-stream framework integrates multidimensional financial-deterioration prediction and prospective compliance-enforcement risk with leakage-controlled temporal validation, probability calibration, explainable risk attribution, and differentiated monitoring prioritization.
Zitieren
APA
Girijala, S., & KV, R. K. (2026). Predictive Risk Analytics for Banking Compliance: An Intelligent Framework for Early Detection of Financial and Operational Risks. Journal of Intelligent Decision Making and Information Science, 3(10s), 242-271. https://doi.org/10.17762/jidmis.v3.2711
Vancouver
Girijala S, KV RK. Predictive Risk Analytics for Banking Compliance: An Intelligent Framework for Early Detection of Financial and Operational Risks. Journal of Intelligent Decision Making and Information Science. 2026;3(10s):242-271. doi:10.17762/jidmis.v3.2711
BibTeX
@article{Girijala2026,
author = {Girijala, Srikanth and KV, Rajesh Kumar},
title = {Predictive Risk Analytics for Banking Compliance: An Intelligent Framework for Early Detection of Financial and Operational Risks},
journal = {Journal of Intelligent Decision Making and Information Science},
year = {2026},
volume = {3},
number = {10s},
pages = {242--271},
doi = {10.17762/jidmis.v3.2711}
}
EUVIMED