Abstract
Bitcoin is the most widely used cryptocurrency for illegal trade in current darknet markets. Owing to the anonymity of its addresses, even though transaction flows are globally visible, Bitcoin clustering remains one of the most challenging and open problems in illegal Bitcoin transaction analysis. In this article, to resolve this problem, we propose a novel multi-layer heuristic algorithm for Bitcoin clustering, which leverages on-chain transactions as well as off-chain application data in the real world. For this purpose, we first explored the unique characteristics of darknet market ecosystems including their trading systems. By conducting an in-depth analysis of the data manually collected for 11 months, we found that some darknet market review data disclosed transactions containing Bitcoin value and item delivery information. We then identified unique Bitcoin addresses associated with the disclosed information, owned by the same darknet providers. Based on address ownership, more accurate market clusters could be created, which have not previously been identified by other clustering algorithms. According to our experimental results, approximately 31.68% of the darknet market review data matched real Bitcoin transactions, and 122 hidden clusters associated with Silk Road 4 were found. This indicates that the proposed algorithm can complement existing clustering methods and significantly reduce the false negative rate by up to 91.7%.
Showing the abstract — retrieve the full paper via the Exa API.