Cybersecurity of Online Financial Systems using Machine Learning Techniques

With the evolution of technologies and increasing the use of internet banking applications, cyber-attacks are gaining extended dimensions. The developed system aims to use Machine Learning techniques to combat fraud from banking systems. The architecture of the solution described in this paper integrates components such as Java Server, Web Customer, Mobile Client and Python Server, where the Random Forest model is hosted. The system has been shown to improve False Negative Rate in detection, offering security measures by encryption, authentication and secure communication, but also an efficient user interface in terms of usability. Our solution demonstrates efficiency in the detection of fraud in the online financing systems, with future directions involve additional customization based on the user’s behavior, real-time data analysis, but also a logging system based on the characteristics of the mobile device.