Credit risk modeling in Nigerian banking sector: A comparative study of machine learning algorithms

Ucheoma Austie Ehimare *

Michael Okpara University of Agriculture, Statistics, Umudike, Abia State, Nigeria.
 
Research Article
International Journal of Science and Research Archive, 2020, 01(01), 231-235.
Article DOI: 10.30574/ijsra.2020.1.1.0046
Publication history: 
Received on 22 October 2020; revised on 25 November 2020; accepted on 29 November 2020
 
Abstract: 
In Nigeria, economic instability, a high unemployment rate and banking rules mean that credit risk continues to be Very significant and could result in loan defaults. When traditional credit risk models, like logistic regression and credit scoring, are used, the data’s complex structure usually cannot be understood well which keeps the models from being very accurate. This work cheques how using machine learning algorithms of different types can lead to improvement in credit risk modelling in Nigerian banks. The research combines specific borrowers’ details with economic information to compare both ML models and standard methods, looking closely at prediction accuracy, what the algorithms reveal and if they are in line with regulation. 
The report shows how ML works better than traditional methods at predicting credit risk globally, showing clear advantages during volatility and also notes where local studies are insufficient. Important matters related to the quality of data, the transparency of models and harmonisation with local laws are sharply examined. The study assesses different ML algorithms by comparing them with Nigerian loan data and analysing how accurate, useful for testing performance and how quickly they run. With the findings, banks will learn how to use ML in a way that supports compliance without missing out on the benefits of new risk management tools. Recommendations include boosting data systems, making AI understandable, to preparing custom rules for using ML in Nigerian banking. 
 
Keywords: 
Credit Risk; Machine Learning; Economic Instability; Risk Management; Predictive Analysis; Algorithm
 
Full text article in PDF: 
 
Copyright information: 

Copyright © 2020 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0