We introduce a concept of similarity between vertices of directed graphs. Let GA and GB be two directed graphs with, respectively, nA and nB vertices. We define an nB times nA similarity matrix S whose real entry sij expresses how similar vertex j (in GA) is to vertex i (in GB): we say that sij is their similarity score. The similarity matrix can be obtained as the limit of the normalized even iterates of Sk+1 = BSkAT + BTSkA, where A and B are adjacency matrices of the graphs and S0 is a matrix whose entries are all equal to 1. In the special case where GA = GB = G, the matrix S is square and the score sij is the similarity score between the vertices i and j of G. We point out that Kleinberg's "hub and authority" method to identify web-pages relevant to a given query can be viewed as a special case of our definition in the case where one of the graphs has two vertices and a unique directed edge between them. In analogy to Kleinberg, we show that our similarity scores are given by the components of a dominant eigenvector of a nonnegative matrix. Potential applications of our similarity concept are numerous. We illustrate an application for the automatic extraction of synonyms in a monolingual dictionary.
%0 Journal Article
%1 Blondel04
%A Blondel, Vincent D.
%A Gajardo, Anah\'ı
%A Heymans, Maureen
%A Senellart, Pierre
%A Dooren, Paul Van
%C Philadelphia, PA, USA
%D 2004
%I Society for Industrial and Applied Mathematics
%J SIAM Rev.
%K LinearAlgebra graph similarity
%N 4
%P 647--666
%R http://dx.doi.org/10.1137/S0036144502415960
%T A Measure of Similarity between Graph Vertices: Applications to Synonym Extraction and Web Searching
%U http://portal.acm.org/citation.cfm?id=1035533.1035557
%V 46
%X We introduce a concept of similarity between vertices of directed graphs. Let GA and GB be two directed graphs with, respectively, nA and nB vertices. We define an nB times nA similarity matrix S whose real entry sij expresses how similar vertex j (in GA) is to vertex i (in GB): we say that sij is their similarity score. The similarity matrix can be obtained as the limit of the normalized even iterates of Sk+1 = BSkAT + BTSkA, where A and B are adjacency matrices of the graphs and S0 is a matrix whose entries are all equal to 1. In the special case where GA = GB = G, the matrix S is square and the score sij is the similarity score between the vertices i and j of G. We point out that Kleinberg's "hub and authority" method to identify web-pages relevant to a given query can be viewed as a special case of our definition in the case where one of the graphs has two vertices and a unique directed edge between them. In analogy to Kleinberg, we show that our similarity scores are given by the components of a dominant eigenvector of a nonnegative matrix. Potential applications of our similarity concept are numerous. We illustrate an application for the automatic extraction of synonyms in a monolingual dictionary.
@article{Blondel04,
abstract = {We introduce a concept of {similarity} between vertices of directed graphs. Let GA and GB be two directed graphs with, respectively, nA and nB vertices. We define an nB times nA similarity matrix S whose real entry sij expresses how similar vertex j (in GA) is to vertex i (in GB): we say that sij is their similarity score. The similarity matrix can be obtained as the limit of the normalized even iterates of Sk+1 = BSkAT + BTSkA, where A and B are adjacency matrices of the graphs and S0 is a matrix whose entries are all equal to 1. In the special case where GA = GB = G, the matrix S is square and the score sij is the similarity score between the vertices i and j of G. We point out that Kleinberg's "hub and authority" method to identify web-pages relevant to a given query can be viewed as a special case of our definition in the case where one of the graphs has two vertices and a unique directed edge between them. In analogy to Kleinberg, we show that our similarity scores are given by the components of a dominant eigenvector of a nonnegative matrix. Potential applications of our similarity concept are numerous. We illustrate an application for the automatic extraction of synonyms in a monolingual dictionary.},
added-at = {2008-09-24T15:07:28.000+0200},
address = {Philadelphia, PA, USA},
author = {Blondel, Vincent D. and Gajardo, Anah\'{\i} and Heymans, Maureen and Senellart, Pierre and Dooren, Paul Van},
biburl = {https://www.bibsonomy.org/bibtex/2fbaef7a3057ff12e16dfd65c42fb0239/mkroell},
description = {A Measure of Similarity between Graph Vertices},
doi = {http://dx.doi.org/10.1137/S0036144502415960},
interhash = {b59d33c99477e70a646615cd0470f459},
intrahash = {fbaef7a3057ff12e16dfd65c42fb0239},
issn = {0036-1445},
journal = {SIAM Rev.},
keywords = {LinearAlgebra graph similarity},
number = 4,
pages = {647--666},
publisher = {Society for Industrial and Applied Mathematics},
timestamp = {2008-09-24T15:07:28.000+0200},
title = {A Measure of Similarity between Graph Vertices: Applications to Synonym Extraction and Web Searching},
url = {http://portal.acm.org/citation.cfm?id=1035533.1035557},
volume = 46,
year = 2004
}