Table 4 Classification results based on different feature subsets using three methods
Model | ML method | Feature subset selection method | Feature number | Classification results (%) | |||
|---|---|---|---|---|---|---|---|
SE | SP | Acc | Gm | ||||
miPlantPre | NBC | PCA | 76 | 92.2 | 92.6 | 92.4 | 92.4 |
CFS | 20 | 93.9 | 97.8 | 95.8 | 95.8 | ||
B-SVM-RFE | 47 | 93.8 | 98.6 | 96.2 | 96.2 | ||
All features | 152 | 92.9 | 98.0 | 95.4 | 95.4 | ||
RF | PCA | 76 | 93.5 | 95.3 | 94.4 | 94.4 | |
CFS | 20 | 95.0 | 97.6 | 96.3 | 96.3 | ||
B-SVM-RFE | 47 | 95.3 | 97.7 | 96.5 | 96.5 | ||
All features | 152 | 95.3 | 97.7 | 96.5 | 96.5 | ||
SVM | PCA | 76 | 94.9 | 99.2 | 97.0 | 97.0 | |
CFS | 20 | 94.3 | 99.1 | 96.7 | 96.7 | ||
B-SVM-RFE | 47 | 95.5 | 99.1 | 97.2 | 97.2 | ||
All features | 152 | 93.9 | 98.5 | 96.2 | 96.2 | ||
miPlantMat | NBC | PCA | 71 | 88.6 | 82.3 | 85.5 | 85.4 |
CFS | 40 | 93.2 | 74.8 | 83.6 | 83.5 | ||
B-SVM-RFE | 63 | 89.8 | 88.4 | 89.1 | 89.1 | ||
All features | 152 | 91.7 | 79.3 | 85.5 | 85.3 | ||
RF | PCA | 71 | 93.2 | 73.2 | 83.2 | 82.6 | |
CFS | 40 | 89.2 | 89.1 | 89.2 | 89.2 | ||
B-SVM-RFE | 63 | 89.7 | 88.6 | 89.2 | 89.2 | ||
All features | 152 | 86.6 | 84.4 | 85.5 | 85.5 | ||
SVM | PCA | 71 | 88.6 | 84.3 | 86.4 | 86.4 | |
CFS | 40 | 90.6 | 87.5 | 89.1 | 89.1 | ||
B-SVM-RFE | 63 | 92.9 | 88.7 | 90.8 | 90.8 | ||
All features | 152 | 87.1 | 81.6 | 84.4 | 84.4 | ||