You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Stable results of the modularity measure and identified communities
Current Behavior
When importing a .csv file, formatted as a Nodes table (Id, Label columns only) and then appending another .csv file, formatted as na Edges table (source, target, weight columns only) I get lower modularity scores and more communities (the decomposition is stable when repeated).
When I first import the edges table and then append the nodes table (or do not append it at all), I get higher modularity scores with less communities, and, again, the result is stable when I repeat the importing and computation.
The graph is weighted and directed, and modularity options are: randomize - yes, use weights - yes, resolution - 1.0, classes start at - 0.
I have also generated a similar graph from fake data, and the problem remains - the order of importing data files changes modularity scores.
Expected Behavior
Stable results of the modularity measure and identified communities
Current Behavior
When importing a .csv file, formatted as a Nodes table (Id, Label columns only) and then appending another .csv file, formatted as na Edges table (source, target, weight columns only) I get lower modularity scores and more communities (the decomposition is stable when repeated).
When I first import the edges table and then append the nodes table (or do not append it at all), I get higher modularity scores with less communities, and, again, the result is stable when I repeat the importing and computation.
The graph is weighted and directed, and modularity options are: randomize - yes, use weights - yes, resolution - 1.0, classes start at - 0.
I have also generated a similar graph from fake data, and the problem remains - the order of importing data files changes modularity scores.
Attached is the study data (assortNodes.csv is the nodes table, outbreak.csv is the edges table) and the fake data (nodesTest.csv and edgesTest.csv)
assortNodes.csv
edgesTest.csv
nodesTest.csv
outbreak.csv
Possible Solution
Steps to Reproduce
Context
Your Environment
The text was updated successfully, but these errors were encountered: