Table 2 Increased loss with the reduction of networks accentuated by the larger size of the networks (comparison between 2000 and 15, 000 vertices)
From: Coarsening effects on k-partite network classification
# of vertices | Reduction | Accuracy | Precision (macro) | Recall (macro) | F-score (macro) |
|---|---|---|---|---|---|
2000 | 20% | 1.01 ± 0.32% | 0.76 ± 0.51% | 1.08 ± 0.34% | 0.92 ± 0.47% |
80% | 6.20 ± 1.12% | 2.54 ± 1.05% | 6.49 ± 1.15% | 6.12 ± 1.42% | |
15,000 | 20% | 1.50 ± 0.52% | 0.57 ± 0.40% | 1.51 ± 0.53% | 1.69 ± 0.76% |
80% | 10.48 ± 1.73% | 0.63 ± 0.82% | 10.63 ± 1.75% | 11.97 ± 2.24% |