Table 2.

Results comparisons on synthetic data, for three different runs. Mean and standard deviation for different number of added noisy samples and features (ρ).

Method	ρ	mean ± std	ρ	mean ± std
JFSS + RLDA	100	96.1 ± 1.0	200	93.1 ± 1.9
JFSS + SVM	100	94.3 ± 1.5	200	87.2 ± 2.7
JFSS + SR	100	91.5 ± 2.4	200	83.1 ± 4.1
SFS + RLDA	100	90.3 ± 2.3	200	74.1 ± 3.9
SFS + SVM	100	90.1 ± 1.4	200	70.0 ± 2.1
RLDA	100	90.1 ± 2.6	200	67.6 ± 4.4
SVM	100	89.3 ± 2.1	200	63.2 ± 4.3