Table 6.
Performance metrics for linear discriminant analysis against each feature-selection technique.
| Feature Selector | Statistics | Precision | Recall | F1-Score | Accuracy |
|---|---|---|---|---|---|
| iNCA | Minimum | 97.790 | 97.768 | 97.779 | 97.774 |
| Maximum | 98.752 | 98.751 | 98.751 | 98.751 | |
| Average | 98.271 | 98.263 | 98.267 | 98.263 | |
| iChi2 | Minimum | 97.674 | 97.666 | 97.670 | 97.666 |
| Maximum | 98.207 | 98.209 | 98.208 | 98.208 | |
| Average | 97.535 | 97.515 | 97.525 | 97.515 | |
| iMRMR | Minimum | 96.526 | 96.526 | 96.526 | 96.526 |
| Maximum | 97.927 | 97.720 | 97.724 | 97.720 | |
| Average | 97.124 | 97.123 | 97.123 | 97.123 |
iNCA: iterative neighborhood component analysis; iChi2: iterative chi-square; iMRMR: iterative maximum relevance–minimum redundancy.