Abstract:
Financial fraud is a growing problem that poses a significant threat to the banking industry,
the government sector, and the public. In response, financial institutions must continuously
improve their fraud detection systems. While preventive and security measures are put in
place to mitigate financial fraud, criminals persistently adjust and develop new methods to
circumvent fraud prevention systems. This dynamic adaptation poses challenges for quantitative
techniques and predictive models. To address the challenge of unbalanced financial
datasets, this study aims to develop rules to detect fraud transactions and improve accuracy
using Anomaly Reduction Boundary Based Oversampling (ARBBO) method. The performance
of the proposed model is evaluated using various metrics such as accuracy, precision,
recall, f1-score, confusion matrix, and ROC values. The proposed model is compared to
several existing machine learning models such as Random Forest (RF), Decision Tree (DT),
Multi-Layer Perceptron (MLP), K-Nearest Neighbor (KNN), Naive Bayes (NB), and Logistic
Regression (LR) using one benchmark dataset. The experimental results demonstrate that the
classifiers performed better with the resampled data, and the suggested Rule-Based model
with ARBBO in financial fraud detection outperformed then other algorithms by achieving
an accuracy and precision of 0.998 and 0.998, respectively.
