University of Hull CSI

Technical Report CSI-0017

Betweenness centrality and Individual and Hub Importances in Social Network Graphs

K. A. Hawick

Archived: 2016

Abstract

Social interaction networks display a number of complex and emergent phenomena. Social interaction data is now available from a number of sources and available networks vary in size from small localised systems to national and global scaled systems. We explore the use of quantitative graph metrics in characterising individuals and sub-graphs of social networks, and focus in particular on the use of betweenness centrality in identifying highly connected individuals and sub-graphs, based on the number of connections that must pass through them. We explore how social systems break up when the most critical individuals or subgraphs are removed. We present results on the emergent component sizes and number as the whole social system can fragment due to the loss of certain key individuals and sub components. We find that modern online social systems tend to be quite robust agains small number of losses, but speculate on how such networks will evolve in time with the eventual withdrawal over long times of certain individuals and sub groups.

Keywords: social networks; betweenness centrality; components; critical connectivity; ranked nodes; systems failure

Full Document Text: Not yet available.

Citation Information: BiBTeX database for CSI Notes.

BiBTeX reference:

@TechReport{CSI-0017,
        Title = {Betweenness centrality and Individual and Hub Importances in Social Network Graphs},
        Author = {K. A. Hawick},
        Institution = {Computer Science, University of Hull},
        Year = {2016},
        Address = {Cottingham Road, Hull HU6 7RX, UK},
        Month = {September},
        Number = {CSI-0017},
        Type = {CSI},
        Keywords = {social networks; betweenness centrality; components; critical connectivity; ranked nodes; systems failure},
        Owner = {kahawick},
        Timestamp = {2016.11.07}
}


[ CSI Index | CSI BiBTeX ]