Table 5 -.

Comparison between mean Dice Kappa, 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, T2w, PD and FLAIR data in the ADC dataset. Blue color indicates the best performance in terms of SI.

Dataset	SI	DER	OER	ICC	TPR	PPV
Naïve Bayes	0.32±0.27	0.53±0.34	0.82±0.21	0.27	0.23	0.96
Logistic Regression	0.57±0.22	0.32±0.36	0.54±0.14	0.97	0.65	0.57
LDA	0.56±0.23	0.41±0.38	0.46±0.20	0.88	0.48	0.83
QDA	0.36±0.26	0.55±0.36	0.74±0.17	0.44	0.26	0.96
KNN	0.66±0.17	0.18±0.18	0.52±0.18	0.99	0.73	0.65
Decision Trees	0.57±0.18	0.27±0.28	0.58±0.18	0.96	0.58	0.62
Random Forests	0.66±0.17	0.16±0.15	0.53±0.19	0.99	0.73	0.64
Bagging	0.63±0.19	0.21±0.26	0.57±0.03	0.99	0.75	0.58
SVM	0.57±0.24	0.32±0.42	0.54±0.11	0.98	0.66	0.60
AdaBoost	0.63±0.20	0.21±0.24	0.53±0.10	0.98	0.70	0.65