Illicit-transaction-address-detect-PU-Learning

Implementation of the paper [arxiv, inproceedings]:

J. Luo, F. Poursafaei and X. Liu, "Towards Improved Illicit Node Detection with Positive-Unlabelled Learning," 2023 IEEE International Conference on Blockchain and Cryptocurrency (ICBC), Dubai, United Arab Emirates, 2023, pp. 1-5, doi: 10.1109/ICBC56567.2023.10174907.

(Here node means graph node, is a transaction address in a transaction network)

Env

Use the pip install -r requirements.txt command to install all of the Python modules and packages listed.

Data

Two public datasets:

Ethereum Phishing Transaction Network: link

Blockchain Elliptic Data Set: link

Run

Run python data_processor.py to generate the csv files needed for the experiments.

Ethereum Phishing Transaction Network data MulDiGraph.pkl is the example for the input dataset in the code.

Run python main.py to get the experiments results of three models: LR, Elkanoto PU, Bagging Pu.

Can choose other graph node embedding learning model from karateclub (default: Role2Vec)

Citation

Will be happy if this work is useful for your research.

@INPROCEEDINGS{luo_towards_nd_pu_icbc_2023,
  author={Luo, Junliang and Poursafaei, Farimah and Liu, Xue},
  booktitle={2023 IEEE International Conference on Blockchain and Cryptocurrency (ICBC)}, 
  title={Towards Improved Illicit Node Detection with Positive-Unlabelled Learning}, 
  year={2023},
  volume={},
  number={},
  pages={1-5},
  doi={10.1109/ICBC56567.2023.10174907}
}