Fair and Interpretable Credit Risk Modelling with Multi-Agent Architecture and SHAP

Pusadkar, Prajwal Suhas (2025) Fair and Interpretable Credit Risk Modelling with Multi-Agent Architecture and SHAP. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (5MB) | Preview
Preview	PDF (Configuration Manual) Download (2MB) | Preview

Abstract

Concerns continue to arise regarding matters such as fairness, transparency, and accountability while credit scoring and loan approval processes are modernized with the introduction of machine learning (ML) models. This project develops a loan approval system that tackles explainability and fairness with a multi-agent architecture and traditional ML models, integrated with explainable artificial intelligence (XAI) techniques. The system consists of four sequential agents: risk scoring (Agent A), credit limit recommendation (Agent B), loan decisioning (Agent C), and explainability (Agent D). Each agent automates distinct tasks aligned with the workflows within the financial industry, forming a cohesive modular pipeline with integrated interpretability. A model was built and trained on a credit dataset with XGBoost, and explainability was integrated with SHAP to yield global and local insights. To enable bulk application uploads with instantaneous prediction, explanation, and fairness audit retrieval, a frontend was built with Streamlit. The system also measures demographic fairness across gender, education, and marital status. This proves that accurate and explainable credit risk models are possible while simultaneously being fair and balanced, with built-in accountability features. The SHAP-x MAS modular architecture supports future deployment to decision aid systems.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Rifai, Hicham UNSPECIFIED
Subjects:	H Social Sciences > HG Finance Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence H Social Sciences > HG Finance > Credit. Debt. Loans. Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	03 Jul 2026 09:29
Last Modified:	03 Jul 2026 09:29
URI:	https://norma.ncirl.ie/id/eprint/9451

Actions (login required)

View Item