12 Knowledge graphs
This chapter covers:
- An introduction to knowledge graphs and their use
- How to extract entities and relationships from text and create a knowledge graph
- Postprocessing techniques on top of knowledge graphs: semantic networks
- Automatic topic extraction
In this chapter we continue the work, started in the previous one, of decomposing a text into a set of meaningful information and storing it in a graph. Here, we have a clear goal in mind: building a knowledge graph.
In this way we will complete the journey we started 11 chapters ago of managing and processing data using graphs as a core technology and mental model. Knowledge graphs represent the “summit” of what has been discussed throughout the entire book. In previous chapters you have learned how to store and process user–item interaction data for providing recommendations in different shapes and forms, how to deal with transactional data and social networks to fight fraud, and more. Now we will dig deeper into how to extract knowledge out of unstructured data.
This chapter is a bit longer than the others, and very dense. You’ll need to read it as a whole to understand not only how to build a knowledge graph out of textual data, but also how to use it for building advanced services. Through diagrams and concrete examples, I’ve tried to make it easier to read and understand; please look at them carefully as you read to be sure that you grasp the key concepts.