Table 2.
Classification with raw samples.
Method | ReliefF optimal features with raw samples | ||||||||
---|---|---|---|---|---|---|---|---|---|
Breast image categories | True positive | True negative | False positive | False negative | Recall rate | Precision | F1-score | Misclassification rate | Accuracy |
BI-RADS 1 | 67 | 334 | 0 | 0 | 100 | 100 | 100 | 0 | 100 |
BI-RADS 2 | 220 | 191 | 0 | 0 | 100 | 100 | 100 | 0 | 100 |
BI-RADS 3 | 20 | 386 | 1 | 4 | 83.3 | 95.2 | 88.9 | 1.2% | 98.5 |
BI-RADS 4A | 8 | 395 | 3 | 5 | 61.5 | 72.7 | 66.7 | 1.9% | 98.05 |
BI-RADS 4B | 5 | 40 | 2 | 3 | 62.5 | 71.4 | 66.7 | 10% | 90 |
BI-RADS 4C | 21 | 382 | 8 | 0 | 100 | 72.4 | 84 | 1.9% | 98.05 |
BI-RADS 5 | 50 | 360 | 1 | 0 | 100 | 98.0 | 99 | 0.2% | 99.75 |
BI-RADS 6 | 5 | 403 | 0 | 3 | 62.5 | 100 | 76.92 | 0.7% | 99.3 |