Predicting potential miRNA-disease associations based on more reliable negative sample selection

BMC Bioinformatics

Table 2 Performance comparison based on six classical classifiers and fivefold cross-validations

classifier	AUC	AUPR	Precision	Recall	F1-score	Accuracy
KR-NSSM
lightGBM	0.9723	0.9787	0.9678	0.8681	0.9150	0.9196
SVM	0.9701	0.9788	0.9701	0.8799	0.9225	0.9263
RF	0.9699	0.9766	0.9731	0.8608	0.9131	0.9185
LR	0.9763	0.9810	0.9630	0.8751	0.9168	0.9208
XGBoost	0.9595	0.9698	0.9655	0.8554	0.9069	0.9125
MLP	0.9492	0.9642	0.9537	0.8527	0.9001	0.9056
Random selection
lightGBM	0.8719	0.8519	0.8099	0.6853	0.7406	0.7629
SVM	0.8865	0.8489	0.8376	0.8015	0.8189	0.8230
RF	0.8106	0.7347	0.7438	0.4573	0.5599	0.6543
LR	0.8655	0.8451	0.8240	0.7004	0.7570	0.7754
XGBoost	0.7724	0.7280	0.7363	0.4042	0.5187	0.6315
MLP	0.7451	0.7276	0.7575	0.4860	0.5892	0.6669

ISSN: 1471-2105