***** To join INSNA, visit http://www.insna.org *****
Dear All,

In the past several years, I have been using generalized similarity measure for bipartite graphs, whereas two nodes in one partition are similar if they are adjacent to similar nodes in the other partition, and vice versa. The algorithm for calculateing generalized similarities was described by Balázs Kovács in "A generalized model of relational similarity," Social Networks, 32(3), July 2010, pp. 197–211.

I wrote a Python module that uses NetworkX and NumPy to calculate generalized similarities in a connected undirected bipartite graph (e.g., representing words and documents). The module provides function generalized_similarity() that takes a NetworkX graph (naturally, bipartite) and returns two similarity graphs for the nodes in either partition. The module also provides function test_generalized_similarity() that applies the algorithm to the "classical" Southern Women graph.

The module is known to work for connected undirected graphs. I have not tested it for directed or unconnected graphs yet.

The module is available for download and non-commercial use at https://github.com/dzinoviev/generalizedsimilarity. Any feedback would be highly appreciated.

See you at Newport Beach in April!

Dmitry Zinoviev
Professor of Computer Science
Suffolk University, Boston, MA 02114
_____________________________________________________________________ SOCNET is a service of INSNA, the professional association for social network researchers (http://www.insna.org). To unsubscribe, send an email message to [log in to unmask] containing the line UNSUBSCRIBE SOCNET in the body of the message.