Table 9 -.
Comparison between mean Dice Kappa (SI), detection/outline error rate (DER/OER), intra-class correlation (ICC), true positive rate (TPR), and positive prediction value (PPV) values of different classifiers for segmentation of WMHs using T1w, and T2w data – ADC and ADNI1 datasets. Blue color indicates the best performance in terms of SI.
| Dataset | ADC | ADNI1 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Measure | SI | DER | OER | ICC | TPR | PPV | SI | DER | OER | ICC | TPR | PPV |
| Naïve Bayes | 0.25±0.22 | 0.59±0.37 | 0.68±0.41 | 0.37 | 0.18 | 0.79 | 0.43±0.26 | 0.52±0.30 | 0.24±0.21 | 0.51 | 0.33 | 0.87 |
| Logistic | 0.16±0.16 | 0.62±0.70 | 0.71±0.47 | 0.43 | 0.42 | 0.12 | 0.42±0.25 | 0.39±0.50 | 0.11±0.14 | 0.79 | 0.66 | 0.35 |
| LDA | 0.28±0.21 | 0.08±0.23 | 1.08±0.55 | 0.46 | 0.24 | 0.66 | 0.48±0.28 | 0.17±0.34 | 0.25±0.20 | 0.61 | 0.46 | 0.72 |
| QDA | 0.20±0.18 | 0.61±0.32 | 0.80±0.38 | 0.23 | 0.13 | 0.86 | 0.36±0.23 | 0.61±0.35 | 0.20±0.17 | 0.28 | 0.25 | 0.92 |
| KNN | 0.27±0.24 | 0.46±0.59 | 0.54±0.25 | 0.83 | 0.44 | 0.21 | 0.58±0.23 | 0.27±0.37 | 0.28±0.18 | 0.90 | 0.67 | 0.55 |
| Decision | 0.37±0.21 | 0.38±0.45 | 0.59±0.22 | 0.93 | 0.38 | 0.43 | 0.57±0.25 | 0.29±0.27 | 0.24±0.18 | 0.94 | 0.57 | 0.66 |
| Random | 0.45±0.22 | 0.25±0.37 | 0.55±0.27 | 0.93 | 0.56 | 0.41 | 0.65±0.23 | 0.19±0.29 | 0.34±0.19 | 0.95 | 0.67 | 0.70 |
| Bagging | 0.24±0.22 | 0.49±0.65 | 0.60±0.34 | 0.68 | 0.55 | 0.17 | 0.57±0.24 | 0.27±0.42 | 0.03±0.03 | 0.86 | 0.73 | 0.52 |
| SVM | 0.37±0.18 | 0.47±0.51 | 0.71±0.35 | 0.58 | 0.48 | 0.48 | 0.46±0.23 | 0.38±0.47 | 0.16±0.11 | 0.23 | 0.54 | 0.62 |
| AdaBoost | 0.44±0.21 | 0.29±0.47 | 0.57±0.31 | 0.92 | 0.53 | 0.42 | 0.64±0.25 | 0.21±0.34 | 0.26±0.10 | 0.95 | 0.66 | 0.72 |