INHALT
Millions of transactions are made daily worldwide in a highly interconnected network of banks. Several malicious users try by any means to bypass any network security levels and take advantage, resulting in the loss of many millions of dollars. Banks and not only, invest a lot in this sector and identifying unusual activities using data mining techniques is a way to identify any pathologies of the transaction system.
Traditional rules-based fraud detection methods often struggle to keep up with evolving fraud patterns, leading to high false-positive rates and missed fraudulent activity. This dissertation investigates the application of data mining techniques to improve fraud detection in financial transactions. A range of machine-learning approaches will be explored, including supervised models such as decision trees and random forests.
The insights from this research aim to support financial institutions in creating more effective fraud detection systems that can adapt to emerging threats and make them more resilient. By leveraging data mining techniques, the study seeks to improve the accuracy and effectiveness of fraud detection, reducing financial losses while minimizing disruption to legitimate transactions. In addition, the potential of hybrid models that combine multiple machine learning techniques is explored, which will contribute to preventing and, therefore, reducing attackwords in the future. |
|