Table 2 The experimental results on the first category of networks, the quality of the extracted community structures are measured in terms of modularity (Q) and normalised mutual information (NMI).
From: Voting Simulation based Agglomerative Hierarchical Method for Network Community Detection
network | measure | ground truth | FastQ | LPA | LPAm | PPC | Attractor | IsoFdp | proposal |
---|---|---|---|---|---|---|---|---|---|
LFR_1000 | Q | 0.43 | 0.356 | 0.326 | 0.385 | 0.404 | 0.356 | 0.36 | 0.41 |
NMI | 1.00 | 0.671 | 0.752 | 0.89 | 0.924 | 0.902 | 0.941 | 0.925 | |
LFR_5000 | Q | 0.38 | 0.275 | 0.122 | 0.149 | 0.271 | 0.197 | 0.308 | 0.342 |
NMI | 1.00 | 0.345 | 0.304 | 0.368 | 0.501 | 0.536 | 0.649 | 0.776 | |
Dolphin | Q | 0.519 | 0.491 | 0.503 | 0.497 | 0.519 | 0.495 | 0.466 | 0.522 |
NMI | 1.00 | 0.733 | 0.837 | 0.744 | 0.812 | 0.691 | 0.629 | 0.783 | |
Risk map | Q | 0.621 | 0.625 | 0.624 | 0.567 | 0.621 | 0.623 | 0.519 | 0.634 |
NMI | 1.00 | 0.894 | 0.848 | 0.888 | 0.803 | 0.834 | 0.714 | 0.945 | |
Scientists collaboration | Q | 0.739 | 0.749 | 0.681 | 0.587 | 0.751 | 0.707 | 0.62 | 0.739 |
NMI | 1.00 | 0.867 | 0.799 | 0.704 | 0.877 | 0.857 | 0.775 | 0.968 |