Skip to main content

Table 5 Deviance information criteria for independent modeling and multivariate modeling (M-based BYM) of the Bayesian models

From: Impact of socioeconomic inequalities on geographic disparities in cancer incidence: comparison of methods for spatial disease mapping

 

Deviance information criteria

Modeling

α(+βχ ι )

\( \alpha +{\mathbf{\mathcal{V}}}_i\left(+\beta {\chi}_{\iota}\right) \)

\( \alpha +{\mathbf{\mathcal{U}}}_i\left(+\beta {\chi}_{\iota}\right) \)

\( \alpha +{\mathbf{\mathcal{U}}}_i+{\mathbf{\mathcal{V}}}_i\left(+\beta {\chi}_{\iota}\right) \)

M-based BYM

with fixed effects

M-based BYM

with random effects

Without covariate

8141.7

8006.8

7969.5

7961.4

6947.84

7131.05

With covariate

8051.9

7941.5

7895

7900

  
  1. (+βχ ι ) in case of additional covariate - \( \mathbf{\mathcal{U}} \): Autocorrelation - \( \mathbf{\mathcal{V}} \): heterogeneity \( \mathbf{\mathcal{X}} \) Townsend index components