Credit Risk Default Prediction: A comparative analysis of Machine Learning Techniques

Opayemi, Olasubomi Olatayo (2025) Credit Risk Default Prediction: A comparative analysis of Machine Learning Techniques. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (851kB) | Preview
Preview	PDF (Configuration Manual) Download (320kB) | Preview

Abstract

Credit risk assessment is a crucial challenge for financial institutions, with traditional statistical models often struggling to capture complex patterns in large, imbalanced datasets. This study addresses the research question: Which machine learning techniques provide the most accurate and reliable predictions for credit risk default, and how do they compare against traditional statistical models? Using a dataset of 148,670 loan records with 34 features sourced from Kaggle, A Python-based pipeline was designed to preprocess data, engineer features, and compare traditional models (Logistic Regression, Decision Tree, K-Nearest Neighbors) with modern ensemble (Random Forest, XGBoost, LightGBM) and deep learning (Neural Networks) approaches. Performance was evaluated using F1-score, precision, recall, and confusion matrices, prioritizing metrics suitable for imbalanced data. Findings indicate that XGBoost and LightGBM outperformed traditional models due to their ability to model complex, non-linear relationships. Key predictors included credit type, interest rate spread, and loan-to-value ratio, with Logistic Regression highlighting co-applicant credit type as a unique factor. The study offers practical recommendations for financial institutions, advocating XGBoost and LightGBM for accuracy and efficiency, and Logistic Regression for interpretability. Limitations include unquantified class imbalance and limited feature engineering. Proposal for future work was that a cross-dataset validation with explainable AI integration and temporal modeling to improve the capacity and applicability.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Byrne, Brian UNSPECIFIED
Subjects:	H Social Sciences > HG Finance > Credit. Debt. Loans. H Social Sciences > HG Finance > Fintech T Technology > T Technology (General) > Information Technology > Fintech Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions:	School of Computing > Master of Science in FinTech
Depositing User:	Ciara O'Brien
Date Deposited:	20 Aug 2026 11:31
Last Modified:	20 Aug 2026 11:31
URI:	https://norma.ncirl.ie/id/eprint/9578

Actions (login required)

View Item