Skip to content

Graph-based analytics and HGT models to identify suspicious transaction networks (fraufulent transaction in fintech and digital wallets)

Notifications You must be signed in to change notification settings

natgluons/Syndicate-Indication-using-Network-Graph-Analytics

Repository files navigation

Syndicate Indication using Network Graph Analytics

This project was developed for Hackathon Bank Indonesia 2024, utilizing network graph analytics to detect fraudulent activities and syndicate behavior.

Overview

Fraud detection requires identifying hidden connections between entities. This project applies Heterogeneous Graph Transformer (HGT) models to analyze transaction networks, detect anomalies, and uncover syndicate activities.

Features

  • Graph-based fraud detection using network analytics
  • HGT model implementation for analyzing complex relationships
  • Automated pattern recognition to identify suspicious entities

Files

  • HGTmodel_Syndicate-Indication-using-Network-Graph-Analytics.py – HGT model implementation
  • RAFM Working Group Proposal (PDF) – Proposal submitted for Hackathon Bank Indonesia 2024

This project demonstrates the potential of graph AI in financial fraud detection.

About

Graph-based analytics and HGT models to identify suspicious transaction networks (fraufulent transaction in fintech and digital wallets)

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages