Table 4 The voting classifier’s evaluation metrics about six feature sets in female MHD patients
From: Use machine learning to help identify possible sarcopenia cases in maintenance hemodialysis patients
Metric | Top 3 Features | Top 3 Features (P < 0.05) | Top 4 Features | Top 4 Features (P < 0.05) | Top 5 Features | Top 5 Features (P < 0.05) |
|---|---|---|---|---|---|---|
ACCTRSa | 63.27% ± 3.54% | 65.67% ± 3.57% | 65.00% ± 2.69% | 66.73% ± 4.28% | 63.94% ± 2.24% | 73.37% ± 3.25% |
ACCTESb | 74.76% ± 6.75% | 70.95% ± 6.55% | 74.76% ± 6.75% | 74.29% ± 8.57% | 73.81% ± 6.48% | 67.62% ± 9.94% |
AVADc | 14.10% ± 5.34% | 8.20% ± 5.45% | 12.37% ± 4.26% | 12.72% ± 5.77% | 11.32% ± 3.96% | 11.25% ± 8.28% |
Precision | 79.03% ± 6.62% | 78.77% ± 7.13% | 79.03% ± 6.62% | 81.86% ± 7.58% | 78.75% ± 6.44% | 77.68% ± 6.73% |
Sensitivity | 81.54% ± 10.43% | 73.85% ± 11.51% | 81.54% ± 10.43% | 76.15% ± 13.95% | 80.00% ± 10.99% | 66.15% ± 14.68% |
Specificity | 63.75% ± 14.20% | 66.25% ± 14.84% | 63.75% ± 14.20% | 71.25% ± 15.86% | 63.75% ± 14.20% | 70.00% ± 8.29% |
F1 Score | 79.81% ± 6.02% | 75.59% ± 6.20% | 79.81% ± 6.02% | 78.04% ± 8.85% | 78.84% ± 5.94% | 70.89% ± 11.06% |
AUC | 77.21% ± 8.43% | 77.50% ± 8.89% | 77.12% ± 7.53% | 77.69% ± 7.92% | 76.54% ± 7.61% | 72.02% ± 8.75% |