liveBook · Manning

Hub node

As explained in Knowledge Graphs and LLMs in Action

A hub node is a graph node connected to an unusually large number of other nodes.

A hub node is a graph node connected to an unusually large number of other nodes. In ontology path search, hubs can create technically valid but semantically misleading connections, so filtering them improves relevance. section 7.7.4

A practical hub-node filtering workflow is:

  • Exclude shortest paths containing selected hubs, then inspect the reduced result set for remaining overly general entities. section 7.7.4

Zika virus path search

Paths between Zika virus and Infectious neuronitis passed through Infectious process and Inflammation, high-degree nodes that connected many unrelated infectious entities. Filtering reduced 11,185 paths to 9, although Viral disease remained overly general.

How Knowledge Graphs and LLMs in Action covers it

How deep it goes
Explains degree-based detection and path exclusion through Cypher, but stops short of fully resolving remaining semantic noise.
Examples
Uses SNOMED paths linking Zika virus disease with Infectious neuronitis through Infectious process, Inflammation, and other general entities.
What it assumes
Assumes familiarity with graph nodes, relationships, Cypher path queries, and ontology data.
What you can do after
Diagnose misleading graph connections with Neo4j Graph Data Science degree scores and exclude paths containing high-degree nodes.
How it connects
Connects graph centrality, degree, graph projection, shortest paths, and ontology relationships.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Knowledge Graphs and LLMs in Action. The section links point to the parts of the book it draws on.