centrality
Let an agent explain centrality measures and their interpretations, support computing and choosing measures for a network, describe applications, and support learning.
Bundle → Layer → Finding → Questions Filled
4 bundles · 8 layers · 8 findings · 16 questions
Understand What centrality is.
Measures
Measures and definitions.
Measures
Measures.
- How are degree, closeness, betweenness, eigenvector, Katz and PageRank centrality defined, and what does each capture? definition
- Which entry fits a specific measure? action
Interpret
Interpretation.
Interpret
Interpretation.
- How should centrality scores be interpreted, and why do measures disagree? definition
- Is centrality being confused with causal influence? boundary
Compute Computing centrality.
Choose
Choosing a measure.
Choose
Choosing.
- Which centrality measure suits this network and question? action
- How are weighted and directed networks handled? definition
Software
Software and scale.
Software
Software.
- How is centrality computed with network analysis libraries, and how are large graphs handled? provenance
- Which entry fits network analysis software? action
Apply Applications.
Social
Social networks.
Social
Social.
- How is centrality used in social network analysis, epidemiology and organisational studies? provenance
- Which findings are contested? boundary
Systems
Infrastructure and biology.
Systems
Systems.
- How is centrality used in transport, power grids, protein networks and web search? provenance
- Which entry fits a specific application? action
Learn History and teaching.
History
History.
History
History.
- How did centrality measures develop from Bavelas and Freeman to PageRank? provenance
- Which references are standard? provenance
Teach
Teaching.
Teach
Teaching.
- How can centrality be taught with small example networks? action
- Which misconceptions arise? provenance
Classifiers Filled
- Family
- Thing Registry
- Category
- Cross-cutting context
- Entry kind
- thing
- Plane
- XCT
- Domain
- XCT.QTY
- Other names and narrower kinds
- eigenvector centrality, closeness centrality, betweenness centrality, Katz centrality
What it is Filled
A family of measures in network analysis that quantify how important or influential a node is within a graph, including degree centrality counting connections, closeness centrality based on distances, betweenness centrality based on shortest paths passing through a node, eigenvector centrality and Katz centrality based on the importance of neighbours, and PageRank; centrality is used in social network analysis, biology, transport, infrastructure and web search.
Why it exists Filled
Let an agent explain centrality measures and their interpretations, support computing and choosing measures for a network, describe applications, and support learning.
Distinguishing features Filled
- Node-level metric
- Multiple definitions
- Interpretation-dependent
- Computable
What robots and AI may and may not do Filled
Must not
- Present a centrality score as proof that a person is a leader, organiser or suspect.
- Rank people in a social network by centrality and act on them without their knowledge or a lawful basis.
- Compute centrality on personal communications data it is not authorized to process.
- Hide the choice of measure when it changes who comes out on top.
Only with a human decision
- Using centrality rankings to target people for investigation, sanctions or exclusion.
May
- Compute degree, betweenness, closeness and other centrality measures on networks it is authorized to analyse.
- State which centrality measure was used and why, since they rank nodes differently.
- Explain the limits of a centrality score as evidence of real influence.
Moral aspects Filled
- Network centrality is used to single out people in surveillance and policing, often on thin evidence.
- A central position can reflect circumstance, not responsibility.
Who is affected
- People in analysed networks
- Analysts and decision-makers
- Communities under surveillance
Owners Filled
Steward
Nobody owns the concept; the owner of a network dataset answers for each analysis.
Links to other meta-models Filled
parent
- Q17004438 - registry parent class
- Q722659 - registry parent class
related
- graph property - category
- metric - category
- classification - ranking and classifying nodes
- control - influence in networks
What else AI and robots need to interact with it Filled
Identity and identifiers required Filled
- Vercy registry: vr.tr.centrality
- Wikidata: Q5062194 (https://www.wikidata.org/wiki/Q5062194)
Direct properties not applicable Not applicable
Not applicable
Plane XCT: no invented physical properties.
Recognition optional Filled
- Node importance in a graph
- Degree, closeness, betweenness, eigenvector
- Density and clustering describe whole networks
- Not physical; scores on nodes of a graph.
Capabilities and actions required Filled
- explain measures
- support computing and choosing
- describe applications
- support learning
Hazards and failure modes required Filled
- Choosing a measure that does not match the question
- Misinterpreting scores as causal influence
- Computational cost on large graphs
Standards and interfaces required Filled
- Mathematical definitions in network science
- Software library conventions
Context of use required Filled
- Measuring node importance in networks.
- degree centrality
- closeness centrality
- betweenness centrality
- eigenvector, Katz and PageRank centrality
- centrality in weighted and directed networks
Sources Filled
- Wikidata item Q5062194: centrality - identity and sense of the item
- Wikipedia: Centrality - general description of the item
Open questions
- Should each measure be a separate entry?
- How should software be linked?
- How should applications be catalogued?
Machine files
Provenance
thing registry research (pass 2) · unreviewed
Built from: models/things/publications/thing-q5062194/spec.json