Temporal-Amount Snapshot MultiGraph for Ethereum Transaction Tracking

With the wide application of blockchain in financial field, rise various types cybercrimes has brought great challenges to security blockchain. In order better understand this emerging market and explore more efficient countermeasures for effective supervision, it is imperative track transactions on blockchain-based systems. Due openness Ethereum, we can easily access publicly available transaction records, model them as a complex network, further study problem tracking via link prediction, which provides deeper understanding Ethereum from network perspective. Specifically, introduce an embedding based prediction framework that composed temporal-amount snapshot multigraph (TASMG) present walk (TAW). By taking realistic rules features networks into consideration, propose TASMG records then TAW effectively embed accounts their integrates temporal amount information proposed network. Experimental results demonstrate superiority learning informative representations could be method tracking.