Dice / Jaccard / inverse log-weighted similarity between pairs of vertices
It would be nice to be able to calculate the Dice, Jaccard and inverse log-weighted similarity between all pairs of vertices in a graph.
In igraph, those similarity metrics are implemented in the following methods:
It would be particularly nice if the algorithms could take advantage of multiple cores?
I am porting some code from igraph to graph_tool to take advantage of a 64 core machine. So far I get huge speed-ups with graph_tool for metrics that are available in both packages (betweenness, closeness, pagerank), so thanks!