Table 1.
Mean performance of statistical features across all 11 wavelet bases tested
| |
DDSM |
MIAS |
||||
|---|---|---|---|---|---|---|
| Feature § | Sensitivity* (%) | Specificity † (%) | Classification rate ‡ (%) | Sensitivity (%) | Specificity (%) | Classification rate (%) |
| M |
89.2 |
26.6 |
50.3 |
86.8 |
24.1 |
44.4 |
| σ |
94 |
27.6 |
52.8 |
87 |
27.7 |
46.9 |
| S |
90.8 |
29.4 |
52.7 |
91.9 |
20.6 |
43.6 |
| K |
92.8 |
23.7 |
49.8 |
93.5 |
16.8 |
41.6 |
| M + σ |
97.4 |
33.9 |
57.9 |
89.7 |
23.7 |
45 |
| M + S |
97.2 |
38.1 |
60.5 |
91.9 |
25.1 |
46.8 |
| M + K |
96.1 |
35.6 |
58.5 |
94 |
19.6 |
43.6 |
| σ + S |
95.6 |
29.1 |
54.3 |
93.2 |
23.1 |
45.8 |
| σ + K |
96.3 |
28.5 |
54.1 |
94.3 |
18 |
42.6 |
| S + K | 94.2 | 32.2 | 55.7 | 93.9 | 19.5 | 43.6 |
§M - mean, σ- standard deviation, S - skewness, K - kurtosis.
*Sensitivity is defined as TP/(TP + FN).
†Specificity is defined as TN/(TN + FP).
‡Classification rate is defined as (TP + TN)/(TP + TN + FP + FN).