A Study of Bitcoin De-Anonymization: Graph and Multidimensional Data Analysis

Bitcoin was designed to be a decentralized global electronic payment system that does not require verification by a third-party intermediary platform and can be used by anyone originally. Due to its anonymity and globalization, bitcoin has achieved great success and attracted the attention of various illegal traders. In recent years, the number of illegal transactions of bitcoin has been increasing. Although bitcoin can support a certain amount of privacy, the bitcoin users and entity information can be linked by tracking the on-chain information of bitcoin users and combining the public off-chain information. Through bitcoin users de-anonymization, we can obtain some valuable intelligence information, which plays an important role in combating bitcoin-related crimes. In this paper, we build a visual analysis system for bitcoin transactions based on a graph database and use real-world multi-dimensional data sources to analyze the entity information of bitcoin transactions on the chain to achieve the effect of de-anonymization. Besides, we adopt a supervised learning method in our system to predict the legitimacy of unknown bitcoin transactions. Experiments and analyses show that our system can achieve good correlation analysis and de-anonymization. Finally, we put forward the future research direction of the bitcoin de-anonymization field.