In recent years API have been proliferated and increasingly connected as well. It has created a warm and welcoming environment for the fraudsters. But it has resulted in a growth in vulnerabilities especially in the field of computer fraud and security. Fraudsters are showing their hand in all the business across the domains. Organizations looking for a platform which gives best solution across the companies. Machine learning (ML) is an efficient and latest technology to give solution for this problem. In this paper, we use ML, we can able to detect and prevent the fraud attacks in real time. With the legacy rule-based approach, organizations are not able to handle fraud when volume of transactions are extremely high and/or precisely. The ML algorithm can handle these transactions to mitigate the possible attacks in these complex legacy systems. The simulation is conducted to model the efficacy of the ML classification during attacks. The results show that the proposed method has higher rate of classifying the attacks than the existing methods.

Classification, attacks, financial transaction, machine learning