Skip to main content

Table 4 Accuracy results in a 50 observations non-proportional hazards, zero skew, 10 and 30% censoring. Prediction accuracy of all tested approaches when simulated data was generated with 50 observations and the following assumptions: non-proportional hazards, zero skew, 10 and 30% censoring. The table summarizes the mean (and standard deviation) of the following metrics: accuracy, Matthews’ correlation, normalized mutual information (NMI), 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

  10% censoring 30% censoring
Method Accuracy Matthews NMI AUC-ROC Sn Sp F1 Accuracy Matthews NMI AUC-ROC Sn Sp F1
Cox
model
0.59 (0.05) 0.14 (0.15) 0.04 (0.07) 0.55 (0.07) 0.39 (0.05) 0.51 (0.05) 0.40 (0.05) 0.58 (0.06) 0.11 (0.19) 0.07 (0.10) 0.53 (0.07) 0.39 (0.05) 0.51 (0.05) 0.41 (0.04)
Kernel Cox 0.61 (0.08) 0.22 (0.15) 0.15 (0.17) 0.64 (0.07) 0.31 (0.13) 0.79 (0.02) 0.46 (0.13) 0.63 (0.07) 0.24 (0.15) 0.07 (0.09) 0.64 (0.08) 0.31 (0.13) 0.79 (0.02) 0.46 (0.13)
wSVM-KM 0.62 (0.04) 0.08 (0.14) 0.02 (0.03) 0.64 (0.07) 0.29 (0.13) 0.89 (0.02) 0.43 (0.13) 0.59 (0.03) 0.05 (0.15) 0.02 (0.01) 0.64 (0.08) 0.29 (0.13) 0.89 (0.02) 0.43 (0.13)
wSVM-Prop 0.60 (0.04) 0.09 (0.14) 0.02 (0.03) 0.64 (0.07) 0.27 (0.01) 0.82 (0.02) 0.42 (0.14) 0.59 (0.03) 0.05 (0.15) 0.02 (0.01) 0.64 (0.08) 0.27 (0.01) 0.82 (0.02) 0.42 (0.14)
pSVM-linear-KM 0.63 (0.07) 0.23 (0.14) 0.08 (0.08) 0.66 (0.09) 0.72 (0.04) 0.80 (0.06) 0.77 (0.05) 0.61 (0.08) 0.22 (0.17) 0.11 (0.09) 0.66 (0.09) 0.72 (0.03) 0.80 (0.06) 0.78 (0.05)
pSVM-linear-prop 0.61 (0.07) 0.21 (0.14) 0.07 (0.07) 0.65 (0.09) 0.71 (0.04) 0.75 (0.06) 0.71 (0.05) 0.59 (0.09) 0.17 (0.18) 0.11 (0.09) 0.63 (0.09) 0.71 (0.04) 0.74 (0.06) 0.71 (0.05)
pSVM-radial-KM 0.63 (0.04) 0.14 (0.14) 0.04 (0.09) 0.63 (0.08) 0.62 (0.02) 0.82 (0.04) 0.69 (0.17) 0.61 (0.06) 0.17 (0.16) 0.10 (0.21) 0.63 (0.09) 0.62 (0.02) 0.82 (0.04) 0.69 (0.17)
pSVM-radial-prop 0.61 (0.04) 0.14 (0.14) 0.10 (0.21) 0.63 (0.08) 0.59 (0.02) 0.79 (0.05) 0.61 (0.14) 0.56 (0.07) 0.12 (0.14) 0.31 (0.36) 0.63 (0.09) 0.59 (0.03) 0.79 (0.03) 0.61 (0.14)
LUPI-linear-KM 0.62 (0.07) 0.22 (0.15) 0.07 (0.08) 0.63 (0.08) 0.79 (0.04) 0.69 (0.07) 0.71 (0.04) 0.62 (0.07) 0.18 (0.16) 0.03 (0.07) 0.63 (0.09) 0.79 (0.04) 0.69 (0.07) 0.71 (0.04)
LUPI-linear-prop 0.61 (0.07) 0.20 (0.15) 0.07 (0.08) 0.63 (0.08) 0.75 (0.04) 0.62 (0.07) 0.64 (0.04) 0.62 (0.07) 0.18 (0.16) 0.03 (0.07) 0.63 (0.09) 0.75 (0.04) 0.62 (0.06) 0.65 (0.04)
inSVM-gradient 0.66 (0.07) 0.27 (0.14) 0.06 (0.07) 0.67 (0.08) 0.83 (0.03) 0.81 (0.06) 0.81 (0.04) 0.64 (0.07) 0.27 (0.15) 0.10 (0.10) 0.69 (0.09) 0.83 (0.03) 0.81 (0.06) 0.81 (0.04)
inSVM-averaging 0.66 (0.07) 0.28 (0.14) 0.07 (0.07) 0.67 (0.09) 0.81 (0.03) 0.83 (0.06) 0.82 (0.04) 0.65 (0.07) 0.28 (0.16) 0.11 (0.11) 0.68 (0.09) 0.81 (0.03) 0.83 (0.06) 0.83 (0.04)