Part 3 – Making Predictions on a Graph

After building a graph and learning how to characterize and visualize it, it’s time to learn about techniques to make predictions from graph data. You will start by using a well-known Python library, namely scikit-learn, and extract data from Neo4j to build a model. Then, you will learn about node embedding algorithms that are built to automatically create node features based on the graph structure. You will then use these embeddings to build node classification and link prediction pipelines, without the need for a third-party library since everything will be managed by the Graph Data Science (GDS) library.

Finally, in the last chapter, you will build a GDS extension and write your own graph algorithm that behaves in the same way as all built-in GDS procedures.

This part includes the following chapters:

• Chapter 6, Building a Machine Learning Model with Graph Features
• Chapter 7, Automatically Extracting Features...