Skip to main content

Table 5 Table of true positive rate, false positive rate, L2 loss and prediction error for the additive-log-ratio model with SNR of 1.67 and d of 45

From: Bayesian compositional regression with microbiome features via variational inference

Method

\(\rho\)

TPR

FPR

L2

PE

Lasso

0

1 ± 0.000

0.365 ± 0.179

5.265 ± 0.699

0.432 ± 0.308

Lin

0

1 ± 0.000

0.388 ± 0.164

5.310 ± 0.686

0.434 ± 0.331

Bates

0

1 ± 0

0.029 ± 0.058

5.685 ± 0.580

0.358 ± 0.257

Selbal

0

0.770 ± 0.091

0.000 ± 0.000

9.644 ± 0.762

1.810 ± 0.729

VB

0

1 ± 0

0.008 ± 0.015

5.633 ± 0.492

0.378 ± 0.247

Lasso

0.2

1 ± 0

0.334 ± 0.166

5.389 ± 0.623

0.471 ± 0.289

Lin

0.2

1 ± 0

0.308 ± 0.164

5.321 ± 0.686

0.482 ± 0.295

Bates

0.2

1 ± 0

0.022 ± 0.036

5.805 ± 0.454

0.445 ± 0.275

Selbal

0.2

0.667 ± 0

0 ± 0

10.274 ± 0.549

1.364 ± 0.568

VB

0.2

1 ± 0

0.003 ± 0.008

5.705 ± 0.438

0.416 ± 0.263

Lasso

0.4

1 ± 0

0.378 ±0.141

5.384 ± 0.5875

0.523 ± 0.346

Lin

0.4

1 ± 0

0.423 ± 0.141

5.358 ± 0.598

0.536 ± 0.333

Bates

0.4

1 ± 0

0.019 ± 0.035

5.792 ± 0.427

0.417 ± 0.257

Selbal

0.4

0.667 ± 0.024

0 ± 0

9.365 ± 0.547

0.945 ± 0.581

VB

0.4

1 ± 0

0.008 ± 0.015

5.633 ± 0.492

0.441 ± 0.273