Similarity matrices for pairs of graphs

摘要: We introduce a concept of similarity between vertices directed graphs. Let GA and GB be two graphs with respectively nA nB vertices. define × matrix S whose real entry sij expresses how similar vertex i (in GA) is to j GB): we say that their score. In the special case where = G, score G square self-similarity graph G. point out Kleinberg's "hub authority" method identify web-pages relevant given query can viewed as our definition in one has unique edge them. analogy Kleinberg, show scores are by components dominant vector non-negative propose simple iterative compute