Date of Award
2023
Document Type
Thesis
Degree Name
Master of Science in Business Analytics MS-BA
Department
Information Systems, Supply Chain, and Analytics
Committee Chair
Hieu Pham
Committee Member
Yi Tan
Committee Member
Qingyun Zhu
Subject(s)
Computer fraud--Detection, Deep learning (Machine learning), Ethereum (Databases)
Abstract
Detecting fraudulent transactions on the Ethereum network can help cryptocurrency companies that operate on the Ethereum platform protect their users from exposure to fraudsters. The most common fraudulent activities in the cryptocurrency network include phishing and smart Ponzi schemes. Since cryptocurrency technology is still young, most investors lack knowledge of how the smart contacts used in the Ethereum platform operate; hence, they cannot evaluate the risks they are exposed to when carrying out cryptocurrency transactions. The key role that this paper looks at is the application of graph neural networks in the extraction of features of users in the Ethereum platform and their respective transactions to classify them as either fraudulent or not fraudulent. The classification results produced by this research show that application of graph neural networks can be used to detect fraudulent transactions and help managers of the Ethereum platform take the necessary actions toward curbing fraudulent activities.
Recommended Citation
Mwanza, Charity, "Graph neural networks for ethereum fraud detection" (2023). Theses. 449.
https://louis.uah.edu/uah-theses/449