A Systematic Review of Detecting Illicit Bitcoin Transactions

Abstract

In recent years, the application of virtual currency has become a part of people's life. The decentralization and anonymity of Bitcoin have made it a favorite tool for many criminals. Therefore, how to trace illegal activities in Bitcoin transactions has become one of the most important research areas. This paper systematically collects 25 research results in this field since 2018, and divides them into three areas, i.e., supervised learning, unsupervised learning, and topological analysis. The supervised learning method based on machine learning is the current mainstream in this research field. However, we believe that the model can achieve more accurate results after combining unsupervised learning and topological analysis features. Moreover, topology analysis can help to observe the entire or specific part of the Bitcoin trading network from a macro perspective so as to discover the hidden illegal activities. In addition, data visualization techniques can provide structural insights to understand the Bitcoin trading network.