This is our implementation for the NCNMF
We proposed a community detection method based on node centrality under the framework of NMF. Specifically, we designed a new similarity measure which considers the proximity of higher-order neighbors to form a more informative graph regularization mechanism, so as to better refine the detected communities. Besides, we introduced the node centrality and \textit{Gini} impurity to measure the importance of nodes and sparseness of the community memberships, respectively. Then, we proposed a novel sparse regularization mechanism which forces nodes with higher node centrality to have smaller \textit{Gini} impurity.
Python==3.7
Ubuntu server with 3.70-GHz i9-10900K CPU and 128-GB main memory
| Statistic | Texas | Cornell | Washington | Wisconsin | Gene | Citeseer | Reality-call | BZR |
|---|---|---|---|---|---|---|---|---|
| #Nodes | 185 | 195 | 217 | 262 | 1,103 | 3,264 | 6,809 | 14,479 |
| #Edges | 296 | 286 | 404 | 476 | 1,672 | 4,612 | 7,697 | 15,535 |
| #Communities | 5 | 5 | 5 | 5 | 5 | 2 | 6 | 2 |