Table 6 The performance of different classification algorithms on 2-class-dataset according to gose0. Significant values are in bold.
From: Prognosis prediction in traumatic brain injury patients using machine learning algorithms
Algorithm | Acc (%) | Acc rank | Prec (%) | Rec (%) | AUC |
|---|---|---|---|---|---|
NB | 81.67 ± 1.29 | 7 | 52.61% ± 3.42 | 51.71% ± 7.88 | 0.820 ± 0.033 |
RF | 84.45 ± 1.29 | 1 | 76.72% ± 12.75 | 27.89% ± 7.10 | 0.827 ± 0.046 |
KNN(k = 5) | 80.64 ± 2.40 | 10 | 50.11% ± 13.99 | 24.14% ± 7.53 | 0.659 ± 0.048 |
KNN(k = 6) | 81.07 ± 2.43 | 9 | 51.48 ± 13.05 | 24.13 ± 9.87 | 0.679 ± 0.056 |
DT | 82.46 ± 1.15 | 6 | 59.98 ± 6.95 | 30.14 ± 8.62 | 0.703 ± 0.038 |
RI | 83.24 ± 2.94 | 4 | 61.84 ± 12.20 | 39.82 ± 8.62 | 0.797 ± 0.067 |
DL | 81.13 ± 2.77 | 8 | 52.81 ± 8.79 | 55.18 ± 12.29 | 0.845 ± 0.029 |
GBT | 82.82 ± 1.72 | 5 | 55.62 ± 3.88 | 51.72 ± 13.27 | 0.827 ± 0.046 |
LR | 84.03 ± 1.76 | 2 | 64.80 ± 8.94 | 40.08 ± 8.01 | 0.842 ± 0.043 |
GLM | 83.91 ± 2.08 | 3 | 63.39 ± 9.43 | 41.08 ± 6.06 | 0.841 ± 0.039 |
Avg Acc | 82.52 |
