基于动机分析的区块链数字货币异常交易行为识别方法

Due to the chaos in the current cryptocurrency market, blockchain digital currency is used by many malicious traders, leading to a series of abnormal trading behaviors such as "dust" injection, "airdrop" operations, extortion, and scams. Therefore, research on the identification method of abnormal transaction behavior of blockchain digital currency is of great significance for regulating transaction behavior and ensuring cyberspace security. Among the many blockchain digital currencies, the market value of Bitcoin exceeds half of the total market value of all blockchain digital currencies, and is highly representative. Bitcoin is the most successful blockchain application scenario at present and one of the most popular topics in the field of digital currency investment and research in the recent decade. The Bitcoin system has a large number of users, a large transaction scale, and anonymization of addresses, which bring great challenges to the accurate identification of abnormal transaction behavior. So far, many researchers have focused on a particular type of illegal and abnormal trading behavior. But different from their method, given that there is a clear motivation behind any Bitcoin abnormal transaction behavior, this article designs a novel method for identifying Bitcoin's abnormal transaction behavior based on the analysis of transaction motivation. Specifically, we take the two types of abnormal transaction behaviors of airdrop candy and greedy capital injection as typical representatives, and design the two types of abnormal transaction behavior determination rules (i.e. judgment rules for airdrop candy behavior and greed injection behavior), and then abstract the abnormal transaction pattern diagram (i.e. airdrop candy behavior trading pattern and greedy capital injection behavior trading pattern). Based on this, the algorithm for identifying abnormal transaction behaviors of Bitcoin was designed and implemented using subgraph matching technology. In order to evaluate the effectiveness of this method, we collected the historical transaction data of Bitcoin for nearly 30 months, and determined the ground-truth set of abnormal transaction behavior through manual analysis. The experimental results show that the recognition recall rate of airdrop candy behavior is 85.71%, the accuracy is 43.62%, the recognition recall rate of greedy fund injection behavior is 81.25%, and the accuracy is 54.32%. In addition, we focus on the analysis and display of three typical examples of Bitcoin's abnormal transaction behavior (i.e. "dust" injection behavior, WannaCry ransomware, SOXex exchange scam), and further verify the effectiveness of the method proposed in this paper through real cases. At the same time, it also shows that there are many abnormal trading activities in the cryptocurrency market, and the cryptocurrency investment market is constantly being disrupted. Therefore, research to identify Bitcoin's abnormal trading behavior has the potential to provide insights into the wider cryptocurrency ecosystem and the trading behavior of thousands of digital currencies now included. It can also help Bitcoin investors understand the dangers of investing in the market and reduce investment risk in the market. In addition, it is more conducive for national authorities to use the abnormal transaction behavior of cryptocurrencies to regulate investors' investment behavior.