Table 5 Real-life datasets metrics. A 5-fold nested-cross validation approach is applied into real-life datasets. Mean (standard deviation) of 10 resampling datasets is shown. The table summarizes the mean (and standard deviation) of the following metrics: accuracy, area under the Receiver Operating Characteristic curve (AUC-ROC), sensitivity (Sn), specificity (Sp) and F1-score (F1)
From: Enhancing SVM for survival data using local invariances and weighting
| Â | Lung | Stanford2 | PBC | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Method | Accuracy | AUC-ROC | Sn | Sp | F1 | Accuracy | AUC-ROC | Sn | Sp | F1 | Accuracy | AUC-ROC | Sn | Sp | F1 |
Cox model | 0.61 (0.08) | 0.60 (0.08) | 0.52 (0.07) | 0.39 (0.07) | 0.50 (0.08) | 0.61 (0.07) | 0.61 (0.11) | 0.48 (0.07) | 0.42 (0.09) | 0.51 (0.09) | 0.75 (0.11) | 0.85 (0.11) | 0.58 (0.07) | 0.62 (0.09) | 0.68 (0.09) |
Kernel Cox | 0.66 (0.14) | 0.52 (0.03) | 0.60 (0.10) | 0.25 (0.07) | 0.49 (0.13) | 0.51 (0.21) | 0.50 (0.01) | 0.56 (0.12) | 0.31 (0.07) | 0.45 (0.08) | 0.55 (0.11) | 0.50 (0.08) | 0.52 (0.12) | 0.30 (0.07) | 0.55 (0.08) |
wSVM-KM | 0.73 (0.08) | 0.68 (0.16) | 0.63 (0.10) | 0.34 (0.07) | 0.59 (0.10) | 0.59 (0.12) | 0.62 (0.04) | 0.57 (0.11) | 0.29 (0.07) | 0.52 (0.10) | 0.70 (0.09) | 0.75 (0.10) | 0.67 (0.04) | 0.65 (0.11) | 0.39 (0.07) |
wSVM-Prop | 0.70 (0.12) | 0.64 (0.15) | 0.62 (0.12) | 0.33 (0.07) | 0.59 (0.12) | 0.55 (0.09) | 0.59 (0.10) | 0.56 (0.12) | 0.27 (0.07) | 0.51 (0.12) | 0.70 (0.07) | 0.73 (0.05) | 0.59 (0.10) | 0.60 (0.12) | 0.41 (0.09) |
pSVM-linear-KM | 0.89 (0.08) | 0.72 (0.16) | 0.63 (0.10) | 0.23 (0.11) | 0.62 (0.11) | 0.78 (0.07) | 0.63 (0.11) | 0.57 (0.10) | 0.21 (0.11) | 0.58 (0.11) | 0.67 (0.08) | 0.84 (0.08) | 0.65 (0.09) | 0.61 (0.09) | 0.35 (0.12) |
pSVM-linear-prop | 0.88 (0.08) | 0.70 (0.15) | 0.65 (0.16) | 0.36 (0.13) | 0.61 (0.11) | 0.77 (0.07) | 0.63 (0.10) | 0.56 (0.16) | 0.24 (0.13) | 0.57 (0.11) | 0.65 (0.12) | 0.80 (0.10) | 0.68 (0.13) | 0.45 (0.11) | 0.26 (0.13) |
pSVM-radial-KM | 0.82 (0.08) | 0.70 (0.14) | 0.69 (0.09) | 0.55 (0.09) | 0.63 (0.09) | 0.60 (0.10) | 0.61 (0.08) | 0.43 (0.09) | 0.45 (0.12) | 0.50 (0.09) | 0.70 (0.07) | 0.65 (0.13) | 0.65 (0.08) | 0.50 (0.09) | 0.55 (0.12) |
pSVM-radial-prop | 0.81 (0.08) | 0.69 (0.15) | 0.61 (0.11) | 0.55 (0.07) | 0.62 (0.10) | 0.60 (0.11) | 0.59 (0.10) | 0.44 (0.06) | 0.51 (0.11) | 0.52 (0.10) | 0.70 (0.07) | 0.61 (0.10) | 0.63 (0.10) | 0.44 (0.06) | 0.51 (0.07) |
LUPI-linear-KM | 0.92 (0.05) | 0.65 (0.14) | 0.72 (0.14) | 0.67 (0.10) | 0.69 (0.11) | 0.80 (0.07) | 0.63 (0.12) | 0.65 (0.11) | 0.51 (0.10) | 0.61 (0.11) | 0.70 (0.07) | 0.63 (0.09) | 0.63 (0.12) | 0.55 (0.11) | 0.50 (0.10) |
LUPI-linear-prop | 0.92 (0.05) | 0.65 (0.14) | 0.71 (0.11) | 0.61 (0.12) | 0.67 (0.09) | 0.80 (0.07) | 0.63 (0.12) | 0.65 (0.10) | 0.53 (0.13) | 0.58 (0.09) | 0.70 (0.07) | 0.63 (0.09) | 0.61 (0.12) | 0.57 (0.13) | 0.53 (0.13) |
inSVM-gradient | 0.67 (0.08) | 0.68 (0.08) | 0.60 (0.10) | 0.43 (0.12) | 0.58 (0.08) | 0.52 (0.11) | 0.59 (0.10) | 0.49 (0.10) | 0.43 (0.12) | 0.58 (0.08) | 0.68 (0.10) | 0.60 (0.15) | 0.59 (0.11) | 0.52 (0.12) | 0.46 (0.12) |
inSVM-averaging | 0.85 (0.07) | 0.71 (0.13) | 0.76 (0.13) | 0.43 (0.07) | 0.65 (0.12) | 0.78 (0.06) | 0.67 (0.13) | 0.69 (0.13) | 0.56 (0.12) | 0.56 (0.09) | 0.75 (0.02) | 0.74 (0.10) | 0.68 (0.11) | 0.61 (0.12) | 0.57 (0.13) |