Enhancing Fraud Detection in Transactive Energy Markets: Investigating the Impact of Data Augmentation on the Performance of Machine Learning and Deep Learning Techniques

The rise of e-banking has resulted in an increase in fraudulent transactions within Energy Markets, emphasizing the critical need for robust fraud detection systems in financial institutions. A major challenge in fraud detection is the prevalent class imbalance issue found in the datasets used for analysis. Previous research has employed diverse methodologies to develop effective classification models. This study investigates the impact of data augmentation techniques on the performance of both traditional machine learning methods and deep learning methods, aiming to enhance fraud detection capabilities. To accomplish this objective, widely used machine learning techniques, including decision trees, support vector machines, and random forests, are combined with two deep neural network models. Through an extensive series of experiments, the findings demonstrate a significant improvement in the performance of the random forest model when data augmentation is applied. This improvement is measured using the F1-Measure metric and surpasses the performance of the other compared methods. Additionally, the results indicate that, overall, the inclusion of data augmentation enhances model performance in terms of recall but may have an impact on precision. Notably, the performance of deep learning methods is negatively affected by the application of data augmentation. The research contributes to the development of more accurate fraud detection systems in energy markets, benefiting financial institutions and energy market operators in combating fraudulent activities and strengthening transaction security and reliability.