Table 3 Table of true positive rate, false positive rate, L2 loss and prediction error for the additive-log-ratio model with SNR of 0.83 and d of 45
From: Bayesian compositional regression with microbiome features via variational inference
Method | \(\rho\) | TPR | FPR | L2 | PE |
|---|---|---|---|---|---|
Lasso | 0 | 1.000 ± 0.000 | 0.277 ± 0.139 | 10.979 ± 1.114 | 1.766 ± 1.191 |
Lin | 0 | 1.000 ± 0.000 | 0.287 ± 0.140 | 10.978 ± 1.182 | 1.786 ± 1.215 |
Bates | 0 | 0.985 ± 0.053 | 0.026 ± 0.041 | 11.456 ± 1.011 | 1.639 ± 1.113 |
Selbal | 0 | 0.680 ± 0.061 | 0.000 ± 0.000 | 13.925 ± 1.022 | 1.904 ± 1.294 |
VB | 0 | 0.867 ± 0.140 | 0.004 ± 0.009 | 11.784 ± 1.135 | 1.618 ± 1.125 |
Lasso | 0.2 | 0.993 ± 0.033 | 0.314 ± 0.140 | 10.932 ± 1.112 | 1.597 ± 0.884 |
Lin | 0.2 | 0.998 ± 0.017 | 0.329 ± 0.134 | 10.997 ± 1.042 | 1.579 ± 0.869 |
Bates | 0.2 | 0.965 ± 0.080 | 0.042 ± 0.056 | 11.478 ± 0.929 | 1.564 ± 0.929 |
Selbal | 0.2 | 0.672 ± 0.037 | 0.000 ± 0.000 | 14.366 ± 0.903 | 1.746 ± 0.978 |
VB | 0.2 | 0.900 ± 0.113 | 0.002 ± 0.007 | 11.878 ± 0.962 | 1.473 ± 0.889 |
Lasso | 0.4 | 0.965 ± 0.080 | 0.352 ± 0.149 | 10.840 ± 1.248 | 1.692 ± 0.872 |
Lin | 0.4 | 0.972 ± 0.071 | 0.426 ± 0.151 | 10.752 ± 1.280 | 1.674 ± 0.951 |
Bates | 0.4 | 0.868 ± 0.145 | 0.043 ± 0.064 | 11.701 ± 1.241 | 1.602 ± 0.966 |
Selbal | 0.4 | 0.668 ± 0.050 | 0.000 ± 0.000 | 13.395 ± 0.958 | 1.643 ± 0.879 |
VB | 0.4 | 0.802 ± 0.135 | 0.001 ± 0.008 | 12.510 ± 1.064 | 1.413 ± 0.707 |