What is Knowledge Graph?
- Knowledge Graph
- A knowledge graph is a store of entities and the relationships between them, holding facts as connections rather than as documents: a company connected to its founders, its products, its locations and its industry. Google's Knowledge Graph is the best known example, and the major engines maintain equivalents that decide what a name refers to before anything is said about it.
01 In practice
Entry to a knowledge graph is earned by corroboration rather than by assertion. A business declaring facts about itself in schema is one source; the same facts confirmed by independent profiles, registries and coverage are what let an engine hold them with any confidence.
Graph membership is what allows an answer to be assembled at all. An engine that has resolved a business to an entity can combine what the site says with what the graph already holds, whereas an unresolved name leaves the model with prose to interpret and no identity to attach it to.
Absence from a graph shows up as vagueness rather than as an error. An assistant asked about an unresolved business returns something hedged and generic, or attaches facts belonging to a similarly named company, which reads to a buyer as a business nobody knows much about.
02 Related terms
The terms next to this one.
The work behind it
Tell the models what you are.
Markup that tells models what your business is, on pages that render without JavaScript.