zyl199710 / COFRAUD

Correlation-aware fraud detection for IJCAI2023 submission

Don't Ignore Alienation and Marginalization: Correlating Fraud Detection

The open-resourced implementation for IJCAI2023 Submission - COFRAUD.

COFRAUD is a correlation-aware fraud detection model, which innovatively incorporates synergistic camouflage into fraud detection. It achieves significant improvements over state-of-the-art methods.

Environments

COFRAUD framework is implemented on Google Colab and major libraries include:

• Pytorch = 1.13.0 + cu116

• Networkx

• dgl = 0.9.1 + cu116

• numpy

Datasets

All dataset in this paper are from previous works

• Amazon¹ contains 11944 users (9.5% fraudsters) and three types of relation, U-P-U, U-S-U, and U-V-U.

• Yelp² contains 45954 users (14.5% fraudsters) and three types of relation, R-U-R, R-T-R, and R-S-R.

Preliminary

In this paper, we design two statistics to prove the existence of alienation and marginalization of fraudsters.

Please Run:

cd .../COFRAUD

python preprocess_data/alienation.py

python preprocess_data/marginalization.py

Method

The code is being sorted out.

Baseline

• GraphConsis

• CARE-GNN

• PC-GNN

• FRAUDRE

• BWGNN

Reference

• [1] McAuley J J, Leskovec J. From amateurs to connoisseurs: modeling the evolution of user expertise through online reviews[C]//Proceedings of the 22nd international conference on World Wide Web. 2013: 897-908.

• [2] Rayana S, Akoglu L. Collective opinion spam detection: Bridging review networks and metadata[C]//Proceedings of the 21th acm sigkdd international conference on knowledge discovery and data mining. 2015: 985-994.