Table 3 Comparison results of different methods*
From: A novel hybrid framework for metabolic pathways prediction based on the graph attention network
Method | Accuracy (%) | Precision (%) | Recall (%) | \(\mathbf {F_1 (\%)}\) |
|---|---|---|---|---|
SVM | 90.21±0.13 | 61.04±0.21 | 51.87±1.40 | 56.08±1.26 |
kNN | 90.96±0.81 | 59.61±3.20 | 62.15±2.80 | 60.85±1.28 |
NB | 81.97±0.61 | 45.06±1.60 | 59.76±1.50 | 51.37±0.88 |
DT | 81.97±0.61 | 45.06±1.60 | 84.56±1.50 | 81.48±0.88 |
RF | 97.89±0.12 | 84.76±0.78 | 84.45±0.68 | 84.60±0.28 |
GCN | 97.61±0.12 | 89.19±0.52 | 93.38±0.44 | 91.17±0.19 |
HFGAT(ours) | 97.19±0.06 | 90.04±0.28 | 94.12±0.16 | 91.97±0.10 |