.png){kind=icon}
.png){kind=icon}
Return
Bitcoin transaction pattern recognition based on semi-supervised learning
.png){kind=icon}
PRE
AI
DOI：10.1016/j.jocs.2023.102055.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In the complex, anonymous and decentralized Bitcoin network, there exist many types of transaction patterns each of which exhibits its own unique structure. It therefore becomes very useful to analyze transaction records to recognize transaction patterns as well as the various roles that traders play in each type of transactions, which can help better understand the operations inside the network. In this paper, we propose a semi-supervised learning based method for the recognition of Bitcoin transaction patterns after defining Bitcoin transaction patterns and describing some of the existing patterns. The proposed method performs transaction pattern recognition following the clustering approach after completing several key tasks including community classifi-cation, transaction structure creation and feature construction. Results of the experiment using real data show that for known transaction patterns, the proposed method can achieve recall rate of at least 70%. Compared to some existing methods, the proposed method can achieve better recognition results by depending not only on the clustering results, but also on the recognition and clustering results of embedded data, which can enhance the effectiveness and the practicality of Bitcoin transaction pattern recognition.
Keywords:
Blockchain
Bitcoin
Transaction pattern
Complex network
Semi-supervised learning
Journal
IF:
18.3
Papers:
3.1K
Citations:
4.0K