Anomaly Detection on Distributed Ledger Using Unsupervised Machine Learning

Abstract

In recent years, blockchain technology has gained widespread attention for its distributed and immutable ledger system that ensures security and transparency. However, the decentralized nature of the blockchain network also presents unique challenges in detecting fraudulent activities, such as money laundering, phishing, and other illicit transactions that may be executed by malicious actors. The traditional detection methods, such as rule-based systems, may not be sufficient to capture the complex and evolving nature of these activities. This paper proposes an anomaly detection approach for significant entity identification within Worldwide Asset Exchange (WAX) blockchain network using unsupervised machine learning. The proposed approach was evaluated by utilizing three detection algorithms (CBLOF, AE, and IF) for assigning an anomaly score to each account in the dataset. The results indicate the presence of potentially fraudulent activities and the effectiveness of the anomaly ranking mechanism in identifying such cases.