Table 6 The performance of models on the fivefold cross-validation
From: Identification of plant vacuole proteins by using graph neural network and contact maps
Model | Acc (%) | F1-score | Sp (%) | Sn (%) | MCC | ROC-AUC | PR-AUC |
|---|---|---|---|---|---|---|---|
GaussianNB | 62.00 | 0.6289 | 58.30 | 65.11 | 0.2355 | 0.6442 | 0.6553 |
LR | 62.75 | 0.6307 | 61.61 | 63.97 | 0.2554 | 0.6782 | 0.6724 |
RF | 60.25 | 0.5835 | 64.62 | 56.20 | 0.2104 | 0.6227 | 0.6165 |
SVM | 62.75 | 0.6002 | 64.57 | 58.46 | 0.2316 | 0.6270 | 0.6527 |
LightGBM | 71.16 | 0.7063 | 73.48 | 69.67 | 0.4327 | 0.7344 | 0.5825 |
GBDT | 58.00 | 0.6689 | 70.75 | 65.62 | 0.3649 | 0.5856 | 0.6127 |
MLP | 58.00 | 0.5730 | 58.26 | 57.32 | 0.1570 | 0.6322 | 0.6372 |
KNN | 57.25 | 0.5659 | 58.89 | 56.12 | 0.1506 | 0.5830 | 0.7388 |
XGBoost | 61.00 | 0.5636 | 63.21 | 54.19 | 0.1757 | 0.5938 | 0.7388 |
GraphIdn | 89.93 | 0.8917 | 89.70 | 90.47 | 0.8020 | 0.9399 | 0.9191 |