Predicting fraudulent card transactions using machine learning methods
Yurii Kryvenchuk, Alina Yamniuk · 2023
The rapid increase in online financial transactions has led to a growing demand for effective and efficient fraud detection systems. The purpose of this work is to assess the performance of machine learning algorithms on the task of fraud detection using relevant evaluation metrics and to compare their effectiveness. Using transaction data from various publicly available datasets, the algorithms are trained and tested under different configurations to determine their optimal parameters and suitability for the task. The evaluation of these models is performed based on their precision, recall, and F1-score metrics. The object of this work is to provide insights into the best practices and potential challenges associated with implementing these machine learning methods for detecting fraudulent transactions The findings of this investigation will contribute to the understanding of the capabilities and limitations of these algorithms for fraud detection, while also shedding light on potential improvements in their performance.