ILODIUBA, CHIEMELIE BERNARD (2026) CUSTOMER CREDIT SCORING SYSTEM USING MACHINE LEARNING AND WEB-BASED DECISION SYSTEM. Other thesis, Godfrey Okoye University, Enugu Nigeria.
![[img]](https://eprints.gouni.edu.ng/style/images/fileicons/text.png) |
Text
Ilodiuba Chiemelie Full Final Report.pdf Download (2MB) |
Abstract
In this research work, an implementation of a machine learning algorithm for customer credit scoring has been done for Bernard Supermarket Enugu, Nigeria, with the aim of solving problems related to inconsistencies, poor record keeping, and poor accountability associated with unstructured, judgmental credit scoring process. The main idea behind this research is to come up with a decision support system based on a web application, which includes a predictable model that will help in making credit granting, approval, and rejection decisions. Literature review shows that credit scoring makes decisions better than judgments, and there is a need for small interpretable credit scoring models for retail credit applications in daily business operations apart from formal banking institutions. The current study was carried out using an agile development approach and consists of a web application having three tiers, which have HTML, CSS, and JavaScript for the front end, PHP for the application layer, MySQL for data storage, and a logistic regression machine learning algorithm written in Python as the classifier. Firstly, the customer dataset was cleaned and normalized, after which it was divided into two parts in the ratio of 80/20 and used for training and testing the classifier respectively. The classifier resulted in 89.21% accuracy and had an equal level of precision, recall, and F1-scores for both classes. Truly, research finds that a small, interpretable machine learning model can provide objective, consistent, and auditable credit decision support even with low-cost hardware. That is why, the system could be considered a reliable baseline solution for supermarket credit management. The platform can potentially be improved through the inclusion of larger datasets, comparative modelling, and wider usage in various settings.
| Item Type: | Thesis (Other) |
|---|---|
| Subjects: | Q Science > Q Science (General) |
| Divisions: | Faculty of Computing And Information Technology (FACIT) |
| Depositing User: | Cynthia Ugwuoti |
| Date Deposited: | 31 Jul 2026 09:09 |
| Last Modified: | 31 Jul 2026 09:09 |
| URI: | http://eprints.gouni.edu.ng/id/eprint/6108 |
Actions (login required)
 |
View Item |
