Table 2 The performances of KNN classifiers based on different kind of features
From: A study of hepatic fibrosis staging methods using diffraction enhanced imaging
Feature set | Feature number | Accuracy | Sensitivity | Specificity | Training time (s) | Prediction speed (obs/s) |
|---|---|---|---|---|---|---|
FO | 14 | 98.60% | 98.85% | 97.63% | 2.2781 | 18,000 |
GLCM | 36 | 99.20% | 99.48% | 97.96% | 5.3947 | 6500 |
GGCM | 15 | 99.20% | 99.42% | 98.47% | 2.7419 | 14,000 |
FO_GLCM | 50 | 99.60% | 99.90% | 98.65% | 7.2376 | 5100 |
FO_GGCM | 29 | 99.60% | 99.66% | 99.25% | 5.3189 | 9400 |
GLCM_GGCM | 51 | 99.60% | 99.76% | 99.10% | 7.354 | 4800 |
FO_GLCM_GGCM | 65 | 99.80% | 99.19% | 99.92% | 8.3886 | 4300 |