Table 5 The performance of the voting classifier and various models that voting classifier combines
From: Use machine learning to help identify possible sarcopenia cases in maintenance hemodialysis patients
Metric | Male | Female | ||||
|---|---|---|---|---|---|---|
LR | Adaboost | LGBM | VCd | SVM | VCd | |
ACCTRSa | 73.18% ± 3.07% | 80.61% ± 3.68% | 89.39% ± 1.92% | 86.59% ± 1.89% | 66.73% ± 4.28% | 66.73% ± 4.28% |
ACCTESb | 75.36% ± 3.37% | 77.86% ± 6.74% | 78.93% ± 6.48% | 80.71% ± 4.29% | 74.29% ± 8.57% | 74.29% ± 8.57% |
AVADc | 4.77% ± 3.87% | 7.55% ± 5.54% | 10.47% ± 6.76% | 6.61% ± 3.63% | 12.72% ± 5.77% | 12.72% ± 5.77% |
Precision | 71.28% ± 10.45% | 75.37% ± 12.04% | 75.39% ± 8.86% | 79.28% ± 9.86% | 81.86% ± 7.58% | 81.86% ± 7.58% |
Sensitivity | 77.50% ± 14.93% | 76.67% ± 12.80% | 76.67% ± 13.84% | 77.50% ± 11.21% | 76.15% ± 13.95% | 76.15% ± 13.95% |
Specificity | 73.75% ± 11.11% | 78.75% ± 13.17% | 80.62% ± 8.59% | 83.12% ± 9.70% | 71.25% ± 15.86% | 71.25% ± 15.86% |
F1 Score | 72.29% ± 6.08% | 74.67% ± 7.84% | 75.25% ± 9.37% | 77.32% ± 5.36% | 78.04% ± 8.85% | 78.04% ± 8.85% |
AUC | 86.20% ± 4.89% | 85.31% ± 6.38% | 85.03% ± 5.82% | 87.40% ± 4.41% | 77.69% ± 7.92% | 77.69% ± 7.92% |