Speaker
Prof.
Naoki Masuda
(University of Tokyo)
Description
The PageRank is a dominant centrality measure for directed
networks. I would like to discuss the utility of an
alternative centrality measure based on the graph Laplacian.
The Laplacian-based centrality value for a node is in fact
equal to the probability that a new type introduced at the
node takes over the entire population in the voter dynamics.
In addition, the Laplacian-based centrality captures
importance of nodes in various other dynamics on networks
including random walk and synchronization. I also explain
its behavior in networks with community structure, its
algebraic characterization, and the relationship to the
Pagerank. Any applications?
