Modelling Credit Default Risk in Microfinance: Machine Learning Approach

Author

Chinodya, Audrey

Title

Modelling Credit Default Risk in Microfinance: Machine Learning Approach

Abstract

Credit default risk is the risk of loss that microfinance institutions face when a borrower
fails to meet their financial obligations, such as repaying a loan or making interest
payments. The goal of this study is to model credit default risk using machine learning
models and to determine which model is best for forecasting credit default risk.
Stepwise logistic regression, least absolute shrinkage and selection operator (LASSO),
neural network, decision trees, and random forest are the models used in this research
study. The study also shows the elements that influence credit risk. The data used was
obtained from a microfinance in Zimbabwe for the period 2018-2022. There were 12
variables and 47000 observations in the data. The model's efficiency was assessed using
the following metrics: accuracy score, recall score, precision score, F1 score, and AUC
value. The research findings highlight the elements that contribute to credit default risk
in microfinance institutions, as well as the efficiency of machine learning models in
forecasting credit default risk. Based on its strong testing performance, F1 and AUC
value. Random Forest is the best model for modelling credit default risk in microfinance
institutions

Date

2023

Publisher

BUSE

Keywords

Credit Default Risk

Modelling

Supervisor

Ms Hlupo