User Tools

Site Tools


algorithms:hits

HITS Centrality

The Hyperlink-Induced Topic Search (HITS) is a link centrality algorithm designed to rate Web pages, but it is applicable to any directed network. Ranks the nodes according to their authority and hub roles in the network.

  • Hubs: nodes with large in-degree that collect information from authorities (compile information).
  • Authorities: nodes with large out-degree, which produce and send information to the hubs (create information).

The network must be directed. Parallel edges are, in principle, allowed and added up. Self-links are meaningless and should not be present.

Input parameter Type Default Description
Iterations*integer100Number of itertions of the algorithm, between 1 and 500.
Property name for hubs*textHubsName of the node property for hub ranking.
Property name for authorities*textAuthoritiesName of the node property for authority ranking.
Link WeighttextNoneLink property that indicate the strength of the connection.

* Required Field

The following picture shows the hub and authority values for a susbset of the internet network, where nodes are webpages and a directed link indicates a hyperlink between two webpages.



References

  • M.E.J. Newman, Networks, an introduction, Oxford University Press, 2010, ISBN: 978-0-19-920665-0, sec. 7.5.
  • Wikipedia: HITS algorithm
algorithms/hits.txt · Last modified: 2019/01/28 17:21 by systems