Deep Learning Techniques in Financial Fraud Detection

Abstract

Nowadays, the rapid development of information technology brings great convenience to people and the financial industry. However, hidden fraud risks emerge at the same time. We can find increasing financial fraud easily. A typical example is credit card fraud, and it will cause the loss of billions of dollars for financial companies and institutions every year. In the process of solving these new types of financial frauds, some traditional statistical and machine learning methods don't perform well. As a result, lots of experts devote themselves to designing some new machine learning methods, and they do have found some great new methods. This paper gives a comprehensive review of new statistics and machine learning methods in detecting financial fraud. The review summarizes and introduces the core idea of these new methods. It provides a good reference source in guiding the detection of financial fraud for both academic and practical fields with useful information on the statistic and machine learning fields.