Skip to main content
. 2021 Feb 5;15:605115. doi: 10.3389/fnins.2021.605115

TABLE 3.

10-fold cross-validation binary mean classification performance for AD against HC using RELM classifier using different feature selection methods.

Feature selection method Performance metrics Mean (%) Standard deviation Max (%) Min (%)
LASSO Accuracy 87.723 0.468 88.663 85.82
Sensitivity 90.93 0.341 91.525 89.50
Specificity 84.52 0.792 85.891 82.14
F-measure 0.883
FSASL Accuracy 76.181 1.069 78.551 73.45
Sensitivity 76.233 1.255 78.839 72.58
Specificity 75.664 1.264 77.868 72.20
F-measure 0.785
LLCFS Accuracy 75.737 1.004 78.690 71.64
Sensitivity 74.205 1.069 77.031 70.11
Specificity 77.881 1.378 81.036 73.64
F-measure 0.817
CFS Accuracy 80.517 1.737 82.86 74.005
Sensitivity 80.035 1.813 82.22 73.25
Specificity 79.16 1.977 81.79 73.084
F-measure 0.867
SVM-RFE Accuracy 68.57 1.186 70.474 65.60
Sensitivity 75.99 1.676 78.832 71.715
Specificity 60.92 1.301 63.426 57.34
F-measure 0.6743