High Impact Factor : 4.396 icon | Submit Manuscript Online icon |

Hybrid Banking Decision Support System

Author(s):

Harsh Chhatwani , PES Modern College of Engineering; Gaurav Amrutkar, PES Modern College of Engineering; Bushara Shaikh, PES Modern College of Engineering; Arya Khopey, PES Modern College of Engineering; Mr. Yogeshchandra Puranik, PES Modern College of Engineering

Keywords:

Banking Decision Support System, Credit Risk, Loan Eligibility, Streaming Data, Online Learning, Concept Drift, Adaptive Machine Learning, Explainable AI, SHAP, LIME, EMI, KYC

Abstract

Banking decision-support systems increasingly use machine learning to assess credit risk, loan eligibility, and repayment behaviour. However, systems trained on static historical datasets may not adequately reflect changes in borrower behaviour and financial conditions. In addition, predictions may communicate a risk label without explaining influential factors or indicating what applicants could review. This paper proposes a Hybrid Banking Decision Support System that combines historical banking data with sequentially replayed observations, concept-drift detection, adaptive model updating, continuous evaluation, and Explainable Artificial Intelligence (XAI). The system is designed to predict credit or loan risk and provide understandable explanations with advisory improvement guidance. Supporting facilities include EMI calculations, fixed- and floating-rate scenarios, moratorium handling, existing obligations, top-up assessment, and KYC document upload and detail extraction. The planned evaluation will use public datasets whose terms permit research use and compare static and adaptive approaches using Accuracy, Precision, Recall, F1-score, ROC-AUC, drift-detection effectiveness, and update efficiency. The proposed architecture uses Python, Django, MySQL, Scikit-learn, XGBoost, River, SHAP, and LIME. The evaluation will test whether adaptive learning maintains or improves robustness under changing data distributions; an improvement in accuracy is not assumed.

Other Details

Paper ID: IJSRDV14I90031
Published in: Volume : 14, Issue : 9
Publication Date: 01/12/2026
Page(s): 49-55

Article Preview

Download Article