Abstract:As an important global geometric quantity, edge betweennesses can
reflect the impact of corresponding edges in the entire network,
which have a very strong practical significance. However, edge
betweenness ignores the relationship between the vertices which
connected indirectly with the edge. This paper first puts forward a
dissimilarity matrix based on a novel edge betweenness, and applies
it to a spectral clustering algorithm and finally, in order to
detect the community structure in complex networks with this
improved algorithm, we use a correction step which is based on the
gravitation force of the community. This algorithm has been tested
on computer-generated networks and typical real world networks.
Compared to GN, PBD, GK and DA, the proposed algorithm possesses
an apparent advantage.
Tingrui Pei;Yuxin Cao;Zhetao Li;Young-june Choi. Spectral Clustering Based on the Edge Betweenness Dissimilarity Matrix for Community Detection[J]. , 2014, 11(11): 3929-3939.
Tingrui Pei;Yuxin Cao;Zhetao Li;Young-june Choi. Spectral Clustering Based on the Edge Betweenness Dissimilarity Matrix for Community Detection. , 2014, 11(11): 3929-3939.