Table 7. Performance of top 5 ternary classification models of all fish for the external validation set.

Model a	AUC	CA	SE_0 b	SE_1 b	SE_2 b
Pub_SVM	0.85	0.67	0.67	0.67	0.68
Est_SVM	0.84	0.66	0.61	0.70	0.62
KR_SVM	0.84	0.65	0.62	0.70	0.60
Pub_KNN	0.80	0.65	0.58	0.66	0.68
FP_SVM	0.81	0.63	0.63	0.66	0.59

^aAP2D: AP2D fingerprint, Est: estate fingerprint, Ext: extended fingerprint, FP: fingerprint, Graph: graphonly fingerprint, KR: Klekota-Roth fingerprint, Maccs: MACCS fingerprint, Pub: PubChem fingerprint, Sub: substructure fingerprint; ANN: artificial neural network, CT: classification tree, kNN: k-nearest neighbor, NB: naïve bayes, RF: random forest, SVM: support vector machine.

^bSE_0: SE value of data labeled class 0, SE_1: SE value of data labeled class 1, SE_2: SE value of data labeled class 2.