Detection of illicit accounts over the Ethereum blockchain
作者：
Highlights：
• We propose a novel approach to detect illicit accounts on the Ethereum blockchain.
• It uses an XGBoost classifier which takes as input 42 features.
• Experiments were carried out on a new data set of 4681 accounts (2179 are illicit).
• Insights are provided on the importance of all involved features.
• We publish the new data set as a benchmark for future work.
摘要
•We propose a novel approach to detect illicit accounts on the Ethereum blockchain.•It uses an XGBoost classifier which takes as input 42 features.•Experiments were carried out on a new data set of 4681 accounts (2179 are illicit).•Insights are provided on the importance of all involved features.•We publish the new data set as a benchmark for future work.
论文关键词：Blockchain,Ethereum,Fraud detection,Machine learning,XGBoost
论文评审过程：Received 1 October 2019, Revised 24 January 2020, Accepted 16 February 2020, Available online 17 February 2020, Version of Record 2 March 2020.
论文官网地址：https://doi.org/10.1016/j.eswa.2020.113318
