Table 3 Complete experimental data for the smallest (2000) and largest (100,000) number of vertices
From: Coarsening effects on k-partite network classification
# of vert. | Reduct. | Accuracy | Precision (macro) | Recall (macro) | F-score (macro) |
|---|---|---|---|---|---|
2000 | 20% | 2.55 ± 0.40% | 0.76 ± 0.14% | 3.10 ± 0.46% | 3.85 ± 0.62% |
35% | 4.07 ± 0.56% | 1.27 ± 0.18% | 4.85 ± 0.63% | 5.73 ± 0.84% | |
50% | 6.22 ± 0.77% | 1.81 ± 0.24% | 7.31 ± 0.86% | 8.31 ± 1.05% | |
60% | 8.69 ± 1.00% | 2.48 ± 0.31% | 10.09 ± 1.10% | 10.89 ± 1.23% | |
80% | 15.86 ± 1.61% | 4.02 ± 0.40% | 18.31 ± 1.82% | 19.32 ± 1.92% | |
100,000 | 20% | 3.25 ± 1.13% | 1.08 ± 0.38% | 3.25 ± 1.13% | 3.92 ± 1.36% |
35% | 4.79 ± 1.16% | 1.55 ± 0.41% | 4.90 ± 1.17% | 5.77 ± 1.45% | |
50% | 11.25 ± 1.80% | 3.24 ± 0.51% | 11.06 ± 1.78% | 14.24 ± 2.27% | |
60% | 12.67 ± 1.93% | 3.59 ± 0.57% | 12.56 ± 1.93% | 15.91 ± 2.41% | |
80% | 19.81 ± 2.58% | 5.24 ± 0.61% | 19.65 ± 2.56% | 23.22 ± 2.97% |