From: A novel hybrid framework for metabolic pathways prediction based on the graph attention network
Methods | Parameter settings |
---|---|
SVM | Gaussian kernel |
kNN | k=5 |
NB | Multinomial NB; Laplace smoothing (\(\alpha\)=1) |
DT | Gini impurity; Minimal samples=2 |
RF | Trees=300; Gini impurity; Depth=60 |
GCN* | Same as  [16] |
HFGAT | d=70; m=2; batch=10; iteration=100 |