Table 4 F1-score of Ensemble GNN models for different number of base classifiers for the IBD dataset
From: Automatic disease prediction from human gut metagenomic data using boosting GraphSAGE
Number of base classifiers | Boosting GraphSAGE | Bagging GraphSAGE | Boosting GCN | Bagging GCN |
|---|---|---|---|---|
5 | 0.93 | 0.91 | 0.86 | 0.91 |
10 | 0.95 | 0.94 | 0.86 | 0.92 |
15 | 0.93 | 0.92 | 0.87 | 0.94 |
20 | 0.91 | 0.92 | 0.89 | 0.93 |
25 | 0.91 | 0.92 | 0.89 | 0.93 |
30 | 0.9 | 0.91 | 0.92 | 0.92 |
35 | 0.92 | 0.9 | 0.91 | 0.9 |
40 | 0.92 | 0.9 | 0.9 | 0.9 |