12 Node classification and link prediction with GNNs

 

This chapter covers

  • Understanding how to use Graph Neural Networks (GNNs) in real scenarios.
  • Building a node classification system based on GNNs for anti-money laundering.
  • Building a link prediction system based on GNNs for movie recommendation and rating.

This chapter explores the application of Graph Neural Networks (GNNs) in two key tasks: node classification and link prediction. These tasks represent fundamental challenges in graph-based machine learning and are central to many real-world applications.

The first part of this chapter explores the application of GNNs for node classification, with a focus on anti-money laundering (AML) applications. By representing financial transactions as a graph, GNNs can be employed to identify suspicious patterns, classify nodes as licit or illicit, and aid in combating financial fraud.

The second part focuses on link prediction in recommendation systems. In this scenario, we will use a GNN-based approach to predict potential user-movie interactions based on the ratings. By learning embeddings for users and movies and leveraging the links between them, we aim to recommend movies to users based on their preferences.

12.1 Node Classification for Anti-money Laundering

12.1.1 Input Data

12.1.2 Graph Processor – Data Preparation

12.1.3 Graph Processor – Homogenous PyG Graph

12.1.4 Encoder-Decoder Architecture

12.1.5 Evaluation and Analysis

12.2 Link Prediction for Movie Recommendation

12.2.1 Input Data

12.2.1 Graph Processor – Data Preparation

12.2.1 Graph Processor – Heterogeneous PyG Graph

12.2.2 Encoder-Decoder Architecture

12.2.3 Evaluation and Analysis

12.3 Summary

12.4 References