Graph Neural Network-Based Bitcoin Transaction Tracking Model
In recent years, with the rapid development of the digital economy, digital currencies such as Bitcoin and Ethereum have become increasingly popular among the public. Tracking and regulating digital currency transactions have become a challenging technology for the healthy development of the digital...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2023-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10156845/ |
| _version_ | 1797796521960275968 |
|---|---|
| author | Zhiyuan Li Enhan He |
| author_facet | Zhiyuan Li Enhan He |
| author_sort | Zhiyuan Li |
| collection | DOAJ |
| description | In recent years, with the rapid development of the digital economy, digital currencies such as Bitcoin and Ethereum have become increasingly popular among the public. Tracking and regulating digital currency transactions have become a challenging technology for the healthy development of the digital economy. That is because blockchain and peer-to-peer networks are the underlying technologies of digital currencies. Blockchain transaction has some new features, such as stronger anonymity and distributed storage. Therefore, it is difficult for the regulatory system to track the transaction relationships among users. Recent studies have shown that the accuracy, time, and space costs of transaction tracking, as well as the trade-offs between them, still need to be improved. In this article, we propose a new blockchain transaction tracking model called BT<sup>2</sup>(Bitcoin Transaction Tracking Model). BT<sup>2</sup> first combines an improved sampling aggregation algorithm with a graph neural network. And then, it exploits an inductive aggregation method to effectively generate rich node embeddings for a small number of unobserved nodes. Next, the node embedding vectors are used to generate edge information among nodes through message-passing functions. Finally, we can use the edge information to obtain the relations among Bitcoin accounts. This paper evaluates the model on the real-world dataset and explores the impact of various parameters, such as network depth and iteration time, etc. From the experimental results, the model’s average AUC (area under the ROC curve) can reach up to 0.93, and the average accuracy is 86%. The numerical results indicate that the performance of BT<sup>2</sup> is better than the state of art methods. |
| first_indexed | 2024-03-13T03:34:20Z |
| format | Article |
| id | doaj.art-546d11a3c6a24d5888dd67414e50d7f6 |
| institution | Directory Open Access Journal |
| issn | 2169-3536 |
| language | English |
| last_indexed | 2024-03-13T03:34:20Z |
| publishDate | 2023-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj.art-546d11a3c6a24d5888dd67414e50d7f62023-06-23T23:00:33ZengIEEEIEEE Access2169-35362023-01-0111621096212010.1109/ACCESS.2023.328802610156845Graph Neural Network-Based Bitcoin Transaction Tracking ModelZhiyuan Li0https://orcid.org/0000-0002-6088-8086Enhan He1School of Computer Science and Telecommunication Engineering, Jiangsu University, Zhenjiang, ChinaSchool of Computer Science and Telecommunication Engineering, Jiangsu University, Zhenjiang, ChinaIn recent years, with the rapid development of the digital economy, digital currencies such as Bitcoin and Ethereum have become increasingly popular among the public. Tracking and regulating digital currency transactions have become a challenging technology for the healthy development of the digital economy. That is because blockchain and peer-to-peer networks are the underlying technologies of digital currencies. Blockchain transaction has some new features, such as stronger anonymity and distributed storage. Therefore, it is difficult for the regulatory system to track the transaction relationships among users. Recent studies have shown that the accuracy, time, and space costs of transaction tracking, as well as the trade-offs between them, still need to be improved. In this article, we propose a new blockchain transaction tracking model called BT<sup>2</sup>(Bitcoin Transaction Tracking Model). BT<sup>2</sup> first combines an improved sampling aggregation algorithm with a graph neural network. And then, it exploits an inductive aggregation method to effectively generate rich node embeddings for a small number of unobserved nodes. Next, the node embedding vectors are used to generate edge information among nodes through message-passing functions. Finally, we can use the edge information to obtain the relations among Bitcoin accounts. This paper evaluates the model on the real-world dataset and explores the impact of various parameters, such as network depth and iteration time, etc. From the experimental results, the model’s average AUC (area under the ROC curve) can reach up to 0.93, and the average accuracy is 86%. The numerical results indicate that the performance of BT<sup>2</sup> is better than the state of art methods.https://ieeexplore.ieee.org/document/10156845/Blockchainbitcoin account anonymitytransaction trackinggraph neural networklink prediction |
| spellingShingle | Zhiyuan Li Enhan He Graph Neural Network-Based Bitcoin Transaction Tracking Model IEEE Access Blockchain bitcoin account anonymity transaction tracking graph neural network link prediction |
| title | Graph Neural Network-Based Bitcoin Transaction Tracking Model |
| title_full | Graph Neural Network-Based Bitcoin Transaction Tracking Model |
| title_fullStr | Graph Neural Network-Based Bitcoin Transaction Tracking Model |
| title_full_unstemmed | Graph Neural Network-Based Bitcoin Transaction Tracking Model |
| title_short | Graph Neural Network-Based Bitcoin Transaction Tracking Model |
| title_sort | graph neural network based bitcoin transaction tracking model |
| topic | Blockchain bitcoin account anonymity transaction tracking graph neural network link prediction |
| url | https://ieeexplore.ieee.org/document/10156845/ |
| work_keys_str_mv | AT zhiyuanli graphneuralnetworkbasedbitcointransactiontrackingmodel AT enhanhe graphneuralnetworkbasedbitcointransactiontrackingmodel |