Malicious Account Classification Using CNN for Ethereum Blockchain’s Accounts

Abstract

The use of cryptocurrencies for transactions has grown over the past few years. Today, cryptocurrency is the most widely used and rapidly expanding currency in the global financial market. This increase can be attributed to blockchain networks, which offer transparent and secure transactions and record cryptocurrency transactions. However, as the volume of transactions increases, fraud also surfaces, resulting in significant losses for the Ethereum account holders involved. Machine learning has been used to address this issue in a previous study; however, the study only provided a limited set of performance metrics. In this study, a CNN-based algorithm is proposed to identify fraudulent accounts in the Ethereum network. The CNN model is applied to a dataset that includes legitimate and fraudulent transactions over the Ethereum network. The results reveal that the CNN-based model successfully identified fraudulent accounts with an accuracy of 98.67%.