Skip to content

AP4549/GNN-SimMule

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

GNN-SimMule: Neural Intelligence Framework

A sophisticated Graph Neural Network (GNN) simulator designed for financial risk analysis, specifically focusing on the detection of "mule rings" and networked fraud patterns.

Overview

GNN-SimMule leverages modern GNN architectures to model financial transactions as a graph topology. Unlike traditional linear analysis, it utilizes Message Passing (GNN-MP) to propagate risk signals through the network, exposing latent connections and surfacing clusters of suspicious activity.

image - Multiple Scenarios image - Intuitive UX, Animated steps for clarity image - Output after GNN processing image

Key Features

  • Neural Intelligence Engine: Proprietary risk propagation logic (Generate, Aggregate, Update).
  • Interactive Graph Visualizer: Real-time rendering of transaction networks with risk-based color grading.
  • Scenario Simulation: Pre-configured analysis templates (e.g., muleRing) for different fraud topologies.
  • Premium Neo-Finance UI: High-fidelity dark-themed interface with glassmorphism and fluid animations.
  • Intelligence Guide: In-app educational framework explaining the mathematical core of GNN operations.

Tech Stack

  • Frontend: React 18, Vite, TypeScript
  • Styling: Tailwind CSS, Shadcn UI, Lucide Icons
  • Motion: Framer Motion
  • Visualization: Recharts, Custom Canvas Visualizers
  • Verification: Vitest, Playwright

Getting Started

Prerequisites

Installation

# Clone the repository
git clone https://github.com/your-repo/gnn-simmule.git
cd gnn-simmule

# Install dependencies
npm install
# OR
bun install

Development

# Start development server
npm run dev
# OR
bun run dev

Scripts

  • npm run dev: Start the development server.
  • npm run build: Build the production bundle.
  • npm run test: Run unit tests with Vitest.
  • npm run lint: Lint the codebase.

Operational Logic

The simulator follows a strict GNN protocol:

  1. GENERATE: Edge signals are computed from source entity risk.
  2. AGGREGATE: Node-level synthesis of incoming risk using permutation-invariant aggregators.
  3. UPDATE: Recalculation of entity risk profiles based on aggregated signals.

© 2026 Neo-Risk Lab - SimMule v1.0.4

About

A sophisticated Graph Neural Network (GNN) simulator designed for financial risk analysis, specifically focusing on the detection of "mule rings" and networked fraud patterns.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages