Contents - Index


NETWORK > CENTRALITY > HUBS AND AUTHORITIES

PURPOSE Calculates the hub and authority centrality scores of a valued network.

DESCRIPTION Hubs and authorities are a natural generalization of eigenvector centrality. There are two scores for each actor, a hub and an authority score. A high hub actor points to many good authories and a  high authority actor receives from many good hubs.  The authority score of a vertex is therefore proportional to the sum of the hub scores of the vertices on the in-coming ties and the hub score is proportional to the authority scores of the vertices on the out-going ties. These values are the same as the singular vectors derived from a single valued decomposition. There are scores associated with all the singular values but the one associated with the largest singular value is usually used as a centrality score.

PARAMETERS
Input network:
Name of file containing network to be analyzed. Data type: Valued Graph 
 
Output Score: (Default = <filename>-HubAuth).
Name of file which will contain hub and authority centrality measures for each vertex.

No. of dimensions: (Default=1)
Number of singularvalues to be calculated. Note that the first is always postive but the rest may be negative and are therefore difficult to interpret.


LOG FILE A table of the hub and authority score for each actor these are normalized so that the sum of squares is one.

TIMING O(N^3).

COMMENTS 

REFERENCES Kleinberg J M (1999) Authorative sources in a hyperlinked environment. Journal of the ACM