First: thanks for bringing that bug to our attention. That one slipped through, possibly when SourceForge's notification mechanism was broken.
To your point: DistanceStatistics.averageDistances() is in fact calculating closeness centrality, not betweenness centrality (the former is what that bug says).
The closeness centrality of a vertex is defined as the average distance (shortest path length) to each other (reachable) vertex. Which is what DistanceStatistics' documentation claims.
It's not impossible that there's a bug in our implementation, but that has yet to be demonstrated.
Do the edges have weights? If the edge weights are < 1, outputs that are all < 1 are not impossible.
Joshua
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thank you for your quick reply and clarification, it gives me confidence to keep plugging away with Java/JUNG rather than running back to my comfort zone: Python/NetworkX.
As to your question: The edges are unweighted, so that's not the problem. You've given me enough information to start tracking down my problem though.
-ells
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I'm trying to calculate the average distance of a graph (~500 vertices), but the output results are uniformly <1. After some googling, I found this (https://sourceforge.net/tracker/?func=detail&aid=3006509&group_id=73840&atid=539119) unresolved bug report from May 2010.
It indicates that it is actually calculating betweenness centrality.
Is JUNG still under active development at this point, or can I expect that this bug wont be resolved?
First: thanks for bringing that bug to our attention. That one slipped through, possibly when SourceForge's notification mechanism was broken.
To your point: DistanceStatistics.averageDistances() is in fact calculating closeness centrality, not betweenness centrality (the former is what that bug says).
The closeness centrality of a vertex is defined as the average distance (shortest path length) to each other (reachable) vertex. Which is what DistanceStatistics' documentation claims.
It's not impossible that there's a bug in our implementation, but that has yet to be demonstrated.
Do the edges have weights? If the edge weights are < 1, outputs that are all < 1 are not impossible.
Joshua
Thank you for your quick reply and clarification, it gives me confidence to keep plugging away with Java/JUNG rather than running back to my comfort zone: Python/NetworkX.
As to your question: The edges are unweighted, so that's not the problem. You've given me enough information to start tracking down my problem though.
-ells