Fit Bayesian Integrated MVN Small Area Estimation model using TMB.

Random effects have a multivariate Gaussian distribution with covariance matrix calculated using integrated_covariance. Kernel hyperparameters are given a prior and learnt.

Usage

ik_tmb(sf, its = 1000, L = 10, type = "hexagonal", ii = NULL, ...)

Arguments

sf

A simple features object with some geometry.

its

Number of iterations in outer loop optimisation, passed to nlminb.

L

The number of Monte Carlo samples to draw from each area.

type

The type argument of sf::st_sample, defaults to "hexagonal"

ii

The (zero-indexed) indices of the observations held-out.

...

Additional arguments to kernel.

Examples

ik_tmb(mw, nsim_warm = 0, nsim_iter = 100, cores = 2)
#> Error in nrow(sf): object 'mw' not found