Study on Machine Learning Techniques with Conventional Tools for Payment Fraud Detection

Abstract

Payment fraud is intentional deception with the purpose of obtaining financial gain or causing loss by implicit or explicit trick, committed by many parties in order to gain significant financial benefits. That had been a major reason for personal financial losses that account over a billion losses a year. At the same time, fraud detection has been improved and currently is embraced by the cutting-edge information technology “Machine Learning”. However, majority of the available studies have been studying with the deep high-end techniques with various costly technologies, and focusing on accuracy and time of the model. They have also been limited to past fraud histories. This study is conducted with multiple machine learning techniques with the use of synthesized dataset, which is not limited to the history, and our study is performed by using the conventional open source data science tools. However, the results seem to be above the expectation.