2.2.1. Parameter Estimation

In the current study, the parameter estimation is done based on a full Bayesian analysis framework, where the appropriate prior distributions have to be assigned to all unknown parameters. In this study, diffuse priors *p*(*β*) ∝ *const* are assigned to all fixed regression parameters and the second-order Gaussian random walk priors were assigned to non-parametric continuous covariates [20]. The structured spatial effects *si* were modeled through a Gaussian Markov random field specified as an intrinsic conditional autoregressive (ICAR) prior distribution [21].
