Financial Fraud Detection in Nigerian Banks: Data Mining Approach
This research investigates the application of data mining techniques, specifically logistic regression and random forest, to detect financial fraud within Nigerian banks. Using individual bank statements and statutory bank charges, the study focuses on developing a robust system for identifying fraudulent transactions. The data preprocessing involves extracting key features such as transaction type, amount, balance, and transaction date. The dataset is split into training and testing sets, and both machine learning models are trained and evaluated based on metrics like accuracy, precision, recall, and F1-scores.The results indicate that the Random Forest model outperforms Logistic Regression, achieving higher accuracy and better handling of complex relationships within the data. Visualization tools like Matplotlib are used to present prediction probabilities, enhancing understanding of model behavior. The system's implementation includes secure access features, detailed transaction analysis, and comprehensive fraud summaries. Challenges such as data imbalance are addressed with techniques like SMOTE and advanced preprocessing methods. This study highlights the potential of using advanced machine learning models for effective fraud detection in financial transactions. The findings suggest that further improvements in feature extraction, data expansion, and exploring more sophisticated models can enhance system performance. This research contributes to the ongoing efforts to secure financial systems against fraudulent activities, offering valuable insights and practical solutions for the banking sector
Authors
- Mohammed Usman
Institutions
- Federal Polytechnic Bali (NG)
Publication Details
- Journal
- INTERNATIONAL JOURNAL OF COMPUTER SCIENCE AND MATHEMATICAL THEORY E-ISSN
- Published
- 2026-11-13
- DOI
- https://doi.org/10.56201/ijcsmt.v10.no3.2024.pg45.66
- Primary Topic
- Imbalanced Data Classification Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00