Table 4 Feature importances of random forest trained on the biggest dataset (\(M=1000\) and \(\max L=100\)) based on normal (a) and LGT (b) network data
From: Constructing phylogenetic networks via cherry picking and machine learning
Features | Importance | |
|---|---|---|
| Â | (a) Normal | (b) LGT |
Leaf distance (t) | 0.190 | 0.162 |
Trivial | 0.155 | 0.184 |
Cherry in tree | 0.143 | 0.146 |
Leaf distance (d) | 0.122 | 0.114 |
LCA distance (t) | 0.068 | 0.056 |
Depth x/y (t) | 0.050 | 0.058 |
Cherry depth (t) | 0.047 | 0.045 |
Depth x/y (d) | 0.043 | 0.038 |
LCA distance (d) | 0.028 | 0.032 |
Leaf depth x (t) | 0.023 | 0.024 |
Leaf depth y (t) | 0.023 | 0.023 |
Cherry depth (d) | 0.020 | 0.023 |
Leaf depth x (d) | 0.020 | 0.022 |
Leaf depth y (d) | 0.020 | 0.022 |
Before/after | 0.015 | 0.016 |
Tree depth (d) | 0.012 | 0.013 |
Tree depth (t) | 0.011 | 0.011 |
New cherries | 0.006 | 0.006 |
Leaves in tree | 0.004 | 0.003 |