Bitcoin, due to its inherent anonymity, creates substantial obstacles for regulatory oversight. This characteristic has led to its widespread use in illicit activities, including money laundering, ransomware collections, and transactions on darknet markets. As a result, tracing the origins of Bitcoin transactions has become increasingly critical. Bitcoin address labeling plays a fundamental role in this process. However, the heterogeneity of data sources, the use of single-dimensional labels, and labeling inaccuracies often lead to conflicts, compromising both the accuracy and completeness of the label data.To address these challenges, this paper proposes a conflict resolution method for Bitcoin address labels based on a multi-dimensional labeling framework. The proposed system first constructs a multi-dimensional label structure, then analyzes and classifies label conflicts arising in various scenarios. It further integrates a truth discovery algorithm with Bitcoin address clustering heuristics to resolve conflicts and identify the most credible labels. Experimental results show that the proposed method effectively resolves multi-label conflicts, thereby enhancing the reliability of Bitcoin address annotations.
Leng et al. (Wed,) studied this question.