A miRNA Target Prediction Model Based on Distributed Representation Learning and Deep Learning

Table 4

Performance evaluation metrics on miRNA target sites prediction.

Dataset	Method	Acc (%)	Sens (%)	Spec (%)	(%)	value
Mock	TargetScan	58.01	60.23	59.22	59.72
	PITA	49.81	58.72	40.82	48.16
	DeepMirTar	93.48	92.35	94.79	93.55
	Our method	96.86	96.97	96.75	96.91	—
Experimented	TargetScan	55.77	39.45	72.08	47.12
	PITA	50.53	13.65	87.41	21.62
	CNN1d	91.05	94.06	87.96	91.40
	Our method	96.04	95.65	96.44	96.09	—

a: 5-fold cross-validation results mean. b: the value according to Student -test (Acc) and indicating comparison between the our model and other methods in experimented test dataset.