Table 5 Optimal set of features at different budget levels for predicting composite ventilation.
From: Budget constrained machine learning for early prediction of adverse outcomes for COVID-19 patients
Budget | Total cost | Optimal set of features | AP mean (std) |
|---|---|---|---|
1 | 1 | DCM, BMP | 0.289 (0.016) |
2 | 2 | DCM, CRP | 0.314 (0.071) |
3 | 3 | DCM, BMP, LDH | 0.331 (0.034) |
4 | 4 | DCM, BMP, CBC, CRP | 0.353 (0.045) |
5 | 4 | DCM, BMP, CBC, CRP | 0.353 (0.045) |
6 | 4 | DCM, BMP, CBC, CRP | 0.353 (0.045) |
7 | 7 | DCM, BMP, CBC, D-dimer, Procalcitonin | 0.363 (0.051) |
8 | 7 | DCM, BMP, CBC, D-dimer, Procalcitonin | 0.363 (0.051) |
9 | 7 | DCM, BMP, CBC, D-dimer, Procalcitonin | 0.363 (0.051) |
10 | 10 | DCM, BMP, CBC, CRP, BNP, Troponin | 0.371 (0.03) |
11 | 10 | DCM, BMP, CBC, CRP, BNP, Troponin | 0.371 (0.03) |
12 | 10 | DCM, BMP, CBC, CRP, BNP, Troponin | 0.371 (0.03) |
13 | 10 | DCM, BMP, CBC, CRP, BNP, Troponin | 0.371 (0.03) |
14 | 10 | DCM, BMP, CBC, CRP, BNP, Troponin | 0.371 (0.03) |
15 | 15 | DCM, BMP, D-dimer, LDH, Sedrate, CRP, BNP, Procalcitonin | 0.382 (0.047) |
16 | 15 | DCM, BMP, D-dimer, LDH, Sedrate, CRP, BNP, Procalcitonin | 0.382 (0.047) |
17 | 15 | DCM, BMP, D-dimer, LDH, Sedrate, CRP, BNP, Procalcitonin | 0.382 (0.047) |
18 | 15 | DCM, BMP, D-dimer, LDH, Sedrate, CRP, BNP, Procalcitonin | 0.382 (0.047) |
19 | 15 | DCM, BMP, D-dimer, LDH, Sedrate, CRP, BNP, Procalcitonin | 0.382 (0.047) |
20 | 15 | DCM, BMP, D-dimer, LDH, Sedrate, CRP, BNP, Procalcitonin | 0.382 (0.047) |
21 | 15 | DCM, BMP, D-dimer, LDH, Sedrate, CRP, BNP, Procalcitonin | 0.382 (0.047) |
22 | 15 | DCM, BMP, D-dimer, LDH, Sedrate, CRP, BNP, Procalcitonin | 0.382 (0.047) |
