An Attention-based Long Short-Term Memory Framework for Detection of Bitcoin Scams

📝 Abstract

A novel Attention-based Long Short-Term Memory (A-LSTM) method is proposed for the classification problem to determine whether a transaction is involved in Ponzi schemes or other cyber scams, or is a non-scam transaction. By integrating an attention mechanism into the LSTM architecture, the model dynamically focuses on the most discriminative temporal patterns in Bitcoin transaction sequences. The attention weights also provide interpretable insights into which transaction behaviors are most indicative of fraudulent activity, making the framework suitable for practical deployment in blockchain fraud detection systems.

📋 BibTeX Citation

@inproceedings{zhao2022bitcoin,
  title     = {An Attention-based Long Short-Term Memory Framework 
               for Detection of Bitcoin Scams},
  author    = {Zhao, Puyang and Tian, W. and Xiao, L. 
               and Liu, X. and Wu, J.},
  booktitle = {2022 International Conference on High Performance 
               Big Data and Intelligent Systems (HDIS)},
  pages     = {21--26},
  year      = {2022},
  month     = {dec},
  publisher = {IEEE},
  address   = {Tianjin, China},
  note      = {Best Paper Nomination}
}

🔗 Related Publications