Introduction
Welcome to the first post of Research Management Insights, a blog dedicated to exploring the intricacies of research project management. In this post, I will provide an overview of my research field and a summary of my research proposal submitted as part of the PROM04 module at the University of Sunderland.
Field of Study
Credit risk modeling is a critical function within the financial services industry, significantly impacting access to loans and broader financial inclusion. This field involves predicting the likelihood that a borrower will default on loan obligations, which is essential for minimizing potential losses and maximizing the profitability of a lender’s portfolio (Bhatore et al., 2020). The evolution from simple linear models to complex machine learning techniques has greatly enhanced the predictive power of credit risk assessments, enabling more nuanced decision-making (Dastile et al., 2020).
Research Proposal Summary
Research Problem: Traditional credit scoring models, while essential for evaluating borrower reliability, often decline in predictive accuracy due to changes in economic conditions and borrower behaviors (Anggodo & Girsang, 2023). This research aims to investigate methodologies that enhance the stability of these models, ensuring consistent performance without compromising their predictive accuracy.
Objectives: The primary objective of this research is to develop more stable and accurate credit risk models. By focusing on both the predictive accuracy and stability of these models, the research seeks to support fairer and more reliable credit assessments, particularly benefiting individuals with limited credit histories (Feldhutter & Schaefer, 2023).
Methodology: The research employs ensemble machine learning techniques, such as Random Forests and Gradient Boosting Machines, known for their robustness and reliability (Tripathi et al., 2021). The study will utilize real-world data provided by Home Credit Group via the Kaggle platform to test the effectiveness of these models (Herman et al., 2024). The Gini Stability Metric will be implemented to quantitatively measure the stability of each model across different economic scenarios and demographic segments (Dorfman, 1979; Řezáč & Řezáč, 2011).
Expected Outcomes: The research aims to contribute to both academic knowledge and practical applications in financial decision-making. By developing more robust models, the project hopes to influence the development of more reliable credit risk management practices, ultimately enhancing decision-making processes within the financial sector and improving risk management (Duman et al., 2023).
Relevance and Impact
Academic Contribution: This research seeks to fill a gap in the current literature regarding the stability of credit risk models. By focusing on model stability in addition to predictive accuracy, the study aims to provide a more comprehensive understanding of credit risk assessment (Pisanie et al., 2023).
Practical Applications: Improved credit risk models will benefit financial institutions by providing more reliable risk assessments, which are crucial for making informed lending decisions. This can lead to better financial inclusion for individuals with limited credit histories and potentially reduce the incidence of loan defaults (Barball et al., 2020).
Conclusion
This research is crucial for developing robust credit scoring models that can adapt to changing conditions. Follow this blog for weekly updates on my progress as I delve deeper into this important field, exploring methodologies, tools, and best practices that drive successful research outcomes.
References
Anggodo, Y. and Girsang, A. (2023) ‘A Novel Modified Binning and Logistics Regression to Handle Shifting in Credit Scoring’. Computational Economics, pp. 1-33.
Barball, J., Loezer, L., Enembreck, F. and Lanzuolo, R. (2020) ‘Lessons learned from data stream classification applied to credit scoring’. Expert Systems with Applications, 162(2020), pp. 1-13.
Bhatore, S., Mohan, L. & Reddy, Y.R. (2023) ‘Machine learning techniques for credit risk evaluation: a systematic literature review’. Journal of Banking and Financial Technology, 4(2020), pp. 111–138.
Cao, Y., Geddes, T., Yang, J. and Yang, P. (2020) ‘Ensemble deep learning in bioinformatics’. Nature Machine Intelligence, 2(2020), pp. 500-508.
Dastile, X., Celik, T. and Potsane, M. (2020) ‘Statistical and machine learning models in credit scoring: A systematic literature survey’. Applied Soft Computing, 91(2020), pp. 1-21.
Dorfman, R. (1979) ‘A Formula for the Gini Coefficient’. The Review of Economics and Statistics, 61(1), pp. 146-149.
Duman, E., Aktas, M. and Yahsi, E. (2023) ‘Credit Risk Prediction Based on Psychometric Data’. Journal of Computers, 12(12), pp. 248-264.
Feldhutter, P. and Schaefer, S. (2023) ‘Debt dynamics and credit risk’. Journal of Financial Economics, 149(2023), pp. 497-535.
Herman, E., McInerney, M., Gaspars-Wieloch, H., Tao, Q., Zhang, Q. (2024) ‘Credit risk assessment using multi-scenario analysis’. Journal of Financial Engineering, 21(3), pp. 102-117.
Pisanie, J., Allison, J. and Visagie, J. (2023) ‘A proposed simulation technique for Population Stability Testing in Credit Risk Scorecards’, Journal of Mathematics, 11(2), pp. 492-508.
Řezáč, M. and Řezáč, F. (2011) ‘How to measure the quality of credit scoring models’. Czech Journal of Economics and Finance, 61(2011), pp. 486-507.
Tripathi, D., Shukla, A., Reddy, B., Bopche, G. and Chandramohan, D. (2021) ‘Credit Scoring Models Using Ensemble Learning and Classification Approaches: A Comprehensive Survey’. Wireless Personal Communications, 123(2022), pp. 785-812.