Graph Machine Learning

Currently, due to the high computational cost of learning, graph scale, data sparsity, noise, and dynamic structures from large networks, Graph Machine learning is essential for modeling complex data. This chapter introduces Graph Machine Learning (GML) as a powerful technique for learning from graph-structured data. It also outlines the main concept of GML and compares it with the traditional machine learning approaches. Supervised and unsupervised tasks such as node classification, link prediction, edge classification, and graph or subgraph classification are discussed to demonstrate how predictive techniques operate on networked data. Recent research in disease prediction and drug discovery shows that GML plays a vital role in healthcare and biomedical research.