ORJI, FRANKLIN CHUKWUEBUKA (2026) A MACHINE LEARNING-BASED LOAN DEFAULT PREDICTION FOR INDIVIDUALS IN RURAL COOPERATIVE SOCIETIES. Other thesis, GODFREY OKOYE UNIVERSITY, ENUGU.
![[img]](https://eprints.gouni.edu.ng/style/images/fileicons/text.png) |
Text
ORJI FRANKLINE COMPLETE WORK.pdf Download (2MB) |
Abstract
The issue of loan default remains a significant challenge among rural cooperatives due to a lack of suitable methods for assessing credit risk, the limited amount of information available about borrowers, and an overreliance on manual methods to make lending decisions. These limitations often result in larger amounts of non-performing loans and associated losses. This study developed a machine learning-based loan default predictive system to help rural cooperatives make data driven decisions about the lending status of individuals applying for credit. The proposed system used the Extreme Gradient Boost (XGBoost) algorithm because of its demonstrated high performance on structured classification problems. The Home Credit dataset provided the borrower and loan data used in this study and the data were cleaned, feature selected, feature engineered, treated for missing values, and encoded prior to training the model. The trained model was incorporated into a web-based application that was built using Flask, React Native, and MongoDB, allowing for real-time assessment of loan risk. The model’s performance was evaluated based 6 on accuracy, confusion matrix, and Receiver Operating Characteristic (ROC) analysis. The model demonstrated an accuracy of 85.4% with an AUC of 0.7569, thus providing evidence of good predictive capability despite class imbalance in the dataset. The system will provide rural cooperatives with an intelligent decision support tool that can assist in identifying high-risk borrowers prior to loan approval; therefore, reducing the risk of defaulting on loans, improving credit management, and promoting sustainable financial activities.
| Item Type: | Thesis (Other) |
|---|---|
| Subjects: | H Social Sciences > H Social Sciences (General) |
| Divisions: | Faculty of Computing And Information Technology (FACIT) |
| Depositing User: | CHIBUEZE EZE |
| Date Deposited: | 24 Jul 2026 14:08 |
| Last Modified: | 24 Jul 2026 14:08 |
| URI: | http://eprints.gouni.edu.ng/id/eprint/5998 |
Actions (login required)
 |
View Item |
