OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Data & Analytics

Dgfraud

DGFraud is an open-source Python toolbox that implements and compares graph neural network models for fraud detection, with example code, datasets and utilities for loading graph data.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Dgfraud — OpenViST project profile in AI Data & Analytics.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use DGFraud when experimenting with GNN-based fraud detection and when relationships among entities can be represented as graph data.

Who it is for

  • Researchers comparing graph-neural-network fraud detectors.
  • Data practitioners preparing labeled graph data for fraud-detection experiments.

Documented uses

  • Inspect and run implemented graph-based fraud-detection models.
  • Compare candidate models using provided example datasets.
  • Prepare graph, feature and label data for a custom model experiment.
Before you choose

Requirements & limits

The public GitHub repository lists an Apache-2.0 license.

  • The repository installation requirements specify Python 3.6 or 3.7 and TensorFlow 1.14 or later but below 2.0.
  • The listed dependencies also include NumPy 1.16.4 or later, SciPy 1.2.0 or later and NetworkX 1.11 or earlier.
  • The README directs users seeking TensorFlow 2.0 to the separate DGFraud-TF2 repository.
  • Custom-data use requires graph adjacency information, features and labels, with feature and label data split into training and testing sets.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Prepare a custom dataset for a DGFraud model

Structure custom graph data for use with a model script.

Before you start

  • A DGFraud model script.
  • Adjacency matrices or, for GAS, adjacency lists.
  • A feature matrix and a label matrix.

Steps

  1. Review load_data_dblp() in utils/utils.py as the repository's example of data loading.

  2. Prepare the graph adjacency data, feature matrix and label matrix.

  3. Split the feature and label matrices into training and testing data.

  4. Set the dataset name with a model script's --dataset argument or edit the script.

Expected outcome

A custom dataset structured for loading by a DGFraud model script.

Guide limitations

  • The README does not specify a universal data schema or preprocessing procedure for every model.
  • Some models require particular graph structures or node types.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
7 November 2026
Original website title
safe-graph/DGFraud · Official homepage title · retained in its source language
How we review sourcesSuggest a correction