Bayesian Hierarchical Compositional Models for Analysing Longitudinal Abundance Data from Microbiome Studies

Table 2

Female dataset at family taxonomic level. Results obtained when estimating the proposed model using different selected principal balances. Additionally, the table shows the deviance information criterion (DIC) defined in [63], the DIC defined in [64], the root mean square deviation (RMSD) of the results obtained fitting the model, and the RMSD of the results obtained using the model for prediction.

Proposed model different SPBal	% var		pD (2002)	DIC (2002)	pD (2004)	DIC (2004)	RMSD (fitting)	RMSD (predicting)
	35.81	−1820.76	16.93	−1803.83	17.60	−1803.15	0.0299	0.0408
,	51.19	−1827.11	24.16	−1802.94	31.55	−1795.56	0.0292	0.0409
, ,	65.27	−1826.61	31.08	−1795.52	40.64	−1785.96	0.0288	0.0442
	75.89	−1832.65	37.68	−1794.96	45.90	−1786.74	0.0282	0.0452
	84.90	−1872.23	45.48	−1826.75	64.49	−1807.74	0.0264	0.0498