Awesome

CommunityEvaluation

A Java framework for evaluating community mining algorithms. Which includes:

• Quality measures for a community structure (e.g. Q-modularity, Variance Ratio Criterion (VRC), Silhouette Width Criterion (SWC), Dunn index, etc.) presented in these paper:
  • R Rabbany et al., Relative Validity Criteria for Community Mining Algorithms; ASONAM 2012
  • R Rabbany et al., Communities validity: methodical evaluation of community mining algorithms; SNAM 2013
  • see measure\criteria
• Agreement indices to compare two given community structures (e.g. Adjusted Rand Index, F-measure, Normalized Mutual Information, etc.), presented in following paper:
  • R Rabbany et al., Generalization of Clustering Agreements and Distances for Overlapping Clusters and Network Communities; CORR 2014
  • see supplementary materials
  • see measure\cluster
  • see a summary and examples in this ipython notebook
• Implementation of the TopLeaders algorithm presented in:
  • R Rabbany et al., Top leaders community detection approach in information networks, SNA-KDD 2010
  • see algorithms\topleaders
• Different graph based distances and centrality measures
  • see measure\graph
• Java wrappers for a collection of community mining algorithms
  • see algorithms\communityMining and execs\
  • most of the executable files would need to be re-build on the specific machine