Mirnacle: machine learning with SMOTE and random forest for improving selectivity in pre-miRNA ab initio prediction

Table 2 Comparison of the selected classifiers combined to the SMOTE filter using TSHS1, TSHS2, and TSHS3 (A) as well as TSMM1, TSMM2, and TSMM3 (B) for the respective stages. Each ML algorithm was tested for each TS through a 10-fold cross validation

Classifier	1st stage			2nd stage			3rd stage
	SN	SL	GM	SN	SL	GM	SN	SL	GM
(A) Results for TSHS1, TSHS2, and TSHS3:
LibSVM	87.5	99.5	93.3	84.7	100	92	98.1	100	99.1
SMO	86.7	84	85.3	92.9	94.5	93.7	99.9	99.4	99.6
MLP	92.3	88.4	90.3	97.9	97	97.5	99.9	99.7	99.8
RF	97.8	97.7	97.8	98.7	99	98.8	100	99.7	99.8
(B) Results for TSMM1, TSMM2, and TSMM3:
LibSVM	89.4	98.8	94	100	60.5	77.8	97.6	100	98.8
SMO	85.6	81.6	83.5	92	92.5	92.2	99.2	99	99.1
MLP	88.9	86.8	87.8	95.2	94.9	95.1	99.9	99.5	99.7
RF	96.4	94.6	95.5	97.8	97.9	97.8	99.8	99.6	99.7

ISSN: 1471-2105