Primary Investigator:
Umit Karabiyik

Project Members
Xiao Hu, Umit Karabiyik

The rapid growth of cryptocurrencies has brought significant challenges to forensic investigations, particularly in detecting trading anomalies. The expanding cryptocurrency ecosystem has created more opportunities for arbitrage between emerging and traditional cryptocurrencies, complicating cross-chain activities. While existing research has primarily focused on price differences across exchanges, little attention has been given to the complexities of cross-chain arbitrage. To address this gap, we propose a blockchain forensics framework that integrates dynamic anomaly detection using the Isolation Forest model and bilateral cross-chain arbitrage analysis, specifically targeting Bitcoin and Dogecoin transactions. By leveraging adaptive statistical modeling, exponential decay factors for dynamic threshold calculation, time-delay analysis, and result visualization, our framework can effectively identify arbitrage opportunities in both directions and uncover irregularities in cross-chain interactions, such as potential fraud, manipulation, or market inefficiencies. Designed to assist forensic investigators, the framework streamlines the process of detecting these anomalies, significantly improving the efficiency and effectiveness of blockchain investigations. Evaluated using real-world transaction datasets, our framework demonstrates its value in advancing blockchain forensic analysis. An open-source implementation of the framework is provided to support reproducibility and facilitate broader application across various blockchain networks.