Table 3.

The performance of ML based models constructed by identified six miRNA signatures on balanced training set and test set using the SMOTE algorithm

Algorithms	Methods	Performance measures
		ACC	Sensitivity	Specificity	MCC	F-score	Precision
SVMR	tenfold	0.990	0.987	0.993	0.981	0.989	0.991
	Test	0.923	0.927	0.919	0.843	0.911	0.895
LR	tenfold	0.688	0.649	0.713	0.357	0.620	0.594
	Test	0.612	0.473	0.716	0.194	0.510	0.553
Naïve Bayes	tenfold	0.761	0.737	0.776	0.508	0.711	0.688
	Test	0.721	0.709	0.730	0.436	0.684	0.661
avNNet	tenfold	0.902	0.918	0.892	0.801	0.882	0.848
	Test	0.783	0.709	0.838	0.553	0.736	0.453
KNN	tenfold	0.889	0.874	0.900	0.773	0.870	0.866
	Test	0.775	0.782	0.770	0.547	0.748	0.717