Statistics: sub-optimal modularity
Bug #727701 reported by
Sébastien Heymann
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Gephi |
Fix Released
|
High
|
Thomas Aynaud |
Bug Description
I have some doubts about the modularity optimisation algorithm that is currently used.
In the wiki, it is stated to be the Louvain algorithm, but the results provided by Gephi often produce lower quality results than the python or matlab code provided online.
for sure, the results can always vary depending on the initial order of encoding of the entries, but even for small graphs with trivial community structure, the algorithm returns suboptimal results.
Does the method implemented in Gephi differ from the original method?
Related branches
lp:~taynaud/gephi/statistics-correction
- Sébastien Heymann: Approve
-
Diff: 163 lines (+14/-39)1 file modifiedStatisticsPlugin/src/org/gephi/statistics/plugin/Modularity.java (+14/-39)
tags: | added: modularity statistics |
Changed in gephi: | |
milestone: | none → 0.7beta |
Changed in gephi: | |
status: | New → Incomplete |
Changed in gephi: | |
status: | Fix Committed → Fix Released |
To post a comment you must log in.
Related question: https:/ /answers. launchpad. net/gephi/ +question/ 147488