Skip to main content
. 2023 Jun 19;128(8):989–998. doi: 10.1007/s11547-023-01657-y

Table 4.

Model of 10 ensembles of random forest classifiers. Classification performance is reported for external testing in terms of accuracy, sensitivity, specificity, PPV and NPV

External testing Random forest (majority vote—38.9% threshold) Radiologist
Accuracy (%) 53 (19/36) 64 (23/36)
Sensitivity (%) 92 (11/12) 88 (10/12)
Specificity (%) 33 (8/24) 54 (13/24)
PPV (%) 41 (11/27) 48 (10/21)
NPV (%) 89 (8/9) 87 (13/15)

p-value = 0.474 (machine learning vs. radiologist)