Bayesian compositional regression with microbiome features via variational inference

BMC Bioinformatics

Table 2 Table of true positive rate, false positive rate, L2 loss and prediction error for the additive-log-ratio model with SNR of 0.5 and d of 200

Method	\(\rho\)	TPR	FPR	L2	PE
Lasso	0	0.820 ± 0.118	0.086 ± 0.056	18.037 ± 3.142	4.673 ± 2.765
Lin	0	0.793 ± 0.130	0.099 ± 0.061	17.830 ± 3.280	4.746 ± 2.753
Bates	0	0.665 ± 0.124	0.007 ± 0.013	19.862 ± 2.434	5.052 ± 3.219
Selbal	0	0.523 ± 0.165	0.001 ± 0.002	21.987 ± 2.050	5.257 ± 3.453
VB	0	0.608 ± 0.138	0.000 ± 0.001	21.392 ± 2.388	4.760 ± 3.139
Lasso	0.2	0.797 ± 0.145	0.097 ± 0.069	18.149 ± 3.537	5.903 ± 3.475
Lin	0.2	0.813 ± 0.136	0.129 ± 0.082	17.236 ± 4.241	6.590 ± 3.841
Bates	0.2	0.662 ± 0.186	0.009 ± 0.015	19.785 ± 2.792	6.229 ± 4.564
Selbal	0.2	0.523 ± 0.162	0.001 ± 0.003	21.988 ± 1.979	6.763 ± 3.814
VB	0.2	0.617 ± 0.141	0.002 ± 0.007	20.739 ± 2.174	6.409 ± 3.599
Lasso	0.4	0.705 ± 0.150	0.112 ± 0.071	17.574 ± 3.755	7.193 ± 4.369
Lin	0.4	0.615 ± 0.203	0.125 ± 0.088	17.864 ± 4.715	8.344 ± 4.456
Bates	0.4	0.545 ± 0.199	0.011 ± 0.015	20.014 ± 2.700	6.401 ± 3.994
Selbal	0.4	0.455 ± 0.162	0.002 ± 0.004	21.533 ± 1.790	6.932 ± 5.012
VB	0.4	0.295 ± 0.180	0.000 ± 0.001	23.043 ± 2.210	9.111 ± 5.407

ISSN: 1471-2105