Fraudulent Account Detection in the Ethereum’s Network Using Various Machine Learning Techniques

Abstract

On the Ethereum network, users communicate with one another through a variety of different accounts. Pseudo-anonymity was enforced over the network to provide the highest level of privacy. By using accounts that engage in fraudulent activity across the network, such privacy may be exploited. Like other cryptocurrencies, Ethereum blockchain may exploited with several fraudulent activities such as Ponzi schemes, phishing, or Initial Coin Offering (ICO) exits, etc. However, the identification of parameters with abnormal account characteristics is not an easy task and requires an intelligent approach to distinguish between normal and fraudulent activities. Therefore, this paper has attempted to solve this a problem by using machine learning techniques to introduce a robust approach that can detect fraudulent accounts on Ethereum. We have used a K-Nearest Neighbor, Random Forest and XGBoost over a collected dataset of 4,681 instances along with 2,179 fraudulent accounts associated and 2,502 regular accounts. The XGBoost, RF, and KNN techniques achieved average accuracies of 96.80 %, 94.8 8%, and 87.85% and an average AUC of 0.995, 0.99 and 0.93, respectively.

Cryptocurrency fraud detection through classification techniques

“…New risks have surfaced in addition to the ease brought about by the quick advancements in information and technology. Attackers have altered their goal, method, and kind of assault due to their fast adaptation to new technology [14], [15]. A new generation of security procedures is required for governmental organizations and institutions, commercial corporations, and ordinary internet users to deal with these risks.…”

An interpretable ensemble model framework for real-time anomaly detection and prediction of Ethereum blockchain transactions

“…Based on the behavior of accounts in Ethereum transactions, [26] investigates the detection of legitimate and unlawful activity. This work developed detection models using a dataset including 4,681 rows of data and three model types (RF, KNN, and XGB).…”

“…Using the XGB model, this investigation achieved an accuracy of 96.3%. Meanwhile, a study by [26,27] also does not use a specific feature selection method and focuses more on manual feature extraction during the data preparation phase for Ethereum account fraud detection.…”

Anomaly detection in smart contracts based on optimal relevance hybrid features analysis in the Ethereum blockchain employing ensemble learning

“…Nonce provides an overview of the quality of transactions between accounts or contracts. At the same time, the Ether Balance reveals the account balance in Wei units, while the hash is a hash code generated by the ethereum virtual machine (EVM), and storage represents the 256-bit hash resulting from the Merkle Root mechanism [19]. The Ethereum network consists of two types of transactions: normal transactions and internal transactions [20].…”