Table 2. IFS performance with SVM and different top features.

Number of features	Sensitivity	Specificity	Accuracy	Matthew correlation coefficient
1	0.938	0.591	0.873	0.560
2	0.875	0.818	0.864	0.620
3	0.885	0.773	0.864	0.603
4	0.885	0.773	0.864	0.603
5	0.885	0.818	0.873	0.636
6	0.865	0.864	0.864	0.638
7	0.917	0.864	0.907	0.723
8	0.927	0.864	0.915	0.743
9	0.927	0.909	0.924	0.775
10	0.938	0.909	0.932	0.796
11	0.948	0.864	0.932	0.785
12	0.927	0.864	0.915	0.743
13	0.927	0.864	0.915	0.743
14	0.927	0.864	0.915	0.743
15	0.927	0.864	0.915	0.743