GNNs: Classifying Graph Nodes with the Spektral Library

Skillsoft issued completion badges are earned based on viewing the percentage required or receiving a passing score when assessment is required. Machine learning (ML) models can be used to extract insights from your graph data. Use this course to learn how to build, train, and evaluate a multi-label classification model using a graph convolutional network (GCN) constructed using the Spektral Python library. Begin by structuring a Spektral dataset for machine learning and learn how data is modeled using an adjacency matrix and feature vectors. Explore how to assign instances of your data to training, validation, and test sets using masks applied to your dataset instance. Construct a graph neural network (GNN) with input layers for the adjacency matrix and features and a GCN convolutional layer and use it to perform node classification. Discover how node features, the edges of the graph, and the structure of the neural network affect the performance of the classification model. Upon completion, you’ll be able to prepare a graph structure for use in an ML model and define the factors which can improve the accuracy of model predictions.