Skip to main content

Table 3 Posterior means and Bayesian credible intervals of predicted relative risk of malaria morbidity estimated from the Bayesian Hierarchical Spatio-temporal model with and without covariates (2016–2021)

From: Modelling spatiotemporal variation in under-five malaria risk in Ghana in 2016–2021

Parameters

Model 1

Model 2

Posterior mean (95% Cr.I)

Posterior mean (95% Cr.I)

Fixed effect

 Intercept

− 0.07 (− 0.12, − 0.26)

10.22 (− 4.1

1, 24.55)

 Log number of children under 5 years

-

0.95 (− 13.82, 15.73)

 Log number of males

− 1.82 (− 16.59, 12.95)

Random effect

\({\tau }_{\mu }^{-1}\)

20,458.78 (1508.20,6.90×10+4)

25,059.51 (1966.54

102,239.83)

\({\tau }_{\vartheta }^{-1}\)

1.13 (1.05, 1.21)

1.49 (1.39, 1.60)

\({\tau }_{\gamma }^{-1}\)

127.70 (22.50, 4.25 ×10+2)

113.54 (21.32, 318.45)

\({\tau }_{\varnothing }^{-1}\)

24,762.67 (1251.56, 1.23 ×10+5)

27,924.84 (2009.64, 122,507.24

Model evaluation metrics

 DIC

21,118.20

21,121.27

 WAIC

20,648.16

20,655.43

  1. Model 1 = Bayesian hierarchical spatio-temporal model without covariates (empty model); Model 2 = Bayesian hierarchical spatio-temporal model with covariates; Cr.I = Credible interval. = Hyperparameter that measures spatial correlation, = Hyperparameter that measures spatial dispersion. \({\tau }_{\gamma }^{-1}\)= Hyperparameter that measures temporal correlation, \({\tau }_{\varnothing }^{-1}\) = Hyperparameter that measures temporal dispersion