Table 8 Table of true positive rate, false positive rate, L2 loss and prediction error for the additive-log-ratio model with SNR of 2.5 and d of 200
From: Bayesian compositional regression with microbiome features via variational inference
Method | \(\rho\) | TPR | FPR | L2 | PE |
|---|---|---|---|---|---|
Lasso | 0 | 1 ± 0 | 0.130 ± 0.077 | 3.384 ± 0.637 | 0.206 ± 0.124 |
Lin | 0 | 1 ± 0 | 0.120 ± 0.064 | 3.412 ± 0.645 | 0.202 ± 0.127 |
Bates | 0 | 1 ± 0 | 0.003 ± 0.007 | 3.847 ± 0.339 | 0.178 ± 0.115 |
Selbal | 0 | 0.667 ± 0.000 | 0.000 ± 0.000 | 9.460 ± 0.367 | 0.396 ± 0.191 |
VB | 0 | 1 ± 0 | 0.001 ± 0.002 | 3.797 ± 0.284 | 0.178 ± 0.106 |
Lasso | 0.2 | 1 ± 0 | 0.150 ± 0.075 | 3.293 ± 0.662 | 0.230 ± 0.138 |
Lin | 0.2 | 1 ± 0 | 0.172 ± 0.069 | 3.107 ± 0.657 | 0.269 ± 0.167 |
Bates | 0.2 | 1 ± 0 | 0.003 ± 0.006 | 3.849 ± 0.302 | 0.184 ± 0.118 |
Selbal | 0.2 | 0.667 ± 0.000 | 0.000 ± 0.000 | 9.399 ± 0.362 | 0.791 ± 0.351 |
VB | 0.2 | 1 ± 0 | 0.001 ± 0.002 | 3.667 ± 0.319 | 0.254 ± 0.156 |
Lasso | 0.4 | 1 ± 0 | 0.170 ± 0.080 | 3.142 ± 0.711 | 0.230 ± 0.133 |
Lin | 0.4 | 1 ± 0 | 0.186 ± 0.067 | 3.095 ± 0.672 | 0.226 ± 0.132 |
Bates | 0.4 | 1 ± 0 | 0.005 ± 0.006 | 3.792 ± 0.313 | 0.180 ± 0.104 |
Selbal | 0.4 | 0.667 ± 0.000 | 0.000 ± 0.000 | 8.019 ± 0.356 | 0.876 ± 0.316 |
VB | 0.4 | 1 ± 0 | 0.000 ± 0.001 | 3.671 ± 0.295 | 0.186 ± 0.110 |