A1870
Title: Efficient fraud detection with AI methods
Authors: Karolina Bolesta - SGH Warsaw School of Economics (Poland) [presenting]
Abstract: Over the last years, the applications of artificial intelligence have grown rapidly. One of the areas in finance where the methods can be leveraged is fraud detection. The purpose is to show what are the most efficient methods to take proactive actions. The focus is on random forest, clustering algorithms and neural networks, showing the possible application use cases. The anomaly detection is presented with the aim to show the most popular unusual patterns and deviations pointing out the fraudulent intent. Based on that, the rules of real-time fraud prevention are explained. This demonstrates how AI methods enable rapid decision-making. The integration and streamlining area is shown to depict the most productive ways for efficient integration of diverse data sources. What is more, the optimization methods are presented for data processing to improve the overall state of fraud detection systems. Demonstrating the challenges in the area and possible ways to overcome them is concluded with.
