st_tgmrf: Spatio-temporal Transformed Gaussian Markov Random Field
In DouglasMesquita/TGMRF: Transformed Gaussian Markov Random Fields

st_tgmrf

R Documentation

Spatio-temporal Transformed Gaussian Markov Random Field

Description

Uses this function to fit spatio-temporal data using non-guassian copulas.

Usage

st_tgmrf(y, X, n_reg, n_var, beta, nu, eps, mu, rho_s, rho_t, rho_st, tau,
  family, type, mat_type, method, nsim, burnin, thin, E, n, neigh,
  prior_param, MCMC_config, fix_rho, range, verbose, c_beta, c_eps, c_mu,
  c_nu, c_rho, conj_beta)

Arguments

`y`	Response variable
`X`	Matrix or data.frame of covariates
`n_reg`	Number of regions in study
`n_var`	Number of variables in study
`beta`	Vector of the initial values to coeficients vector
`nu`	A Initial value to variance of copula
`rho_s`	Spatial dependence parameter
`rho_t`	Temporal dependence parameter
`rho_st`	Spatio-temporal dependence parameter
`tau`	Vector of precisions of variables
`family`	'Poisson' or 'Binary'
`type`	Depends of family. 'lognormal', 'lognormal-precision', 'gamma-shape', 'gamma-scale', 'weibull-shape', 'weibull-scale' for Poisson family 'beta-logit', 'beta-probit', 'beta-alpha', 'beta-beta' for Binomial family
`mat_type`	car or leroux
`method`	'arms' or 'metropolis'
`nsim`	Number of MCMC iterations
`burnin`	number of discards iterations
`thin`	Lag to collect the observations
`E`	Offsets for Poisson data
`n`	Number of trials in Binomial data
`neigh`	Neighborhood structure of nb class or a neighborhood matrix
`prior_param`	List of priors parameters. Gaussian for beta vector and IGamma for nu. Entrys are: 'nu' and 'beta'
`MCMC_config`	Parameters to tunning MCMC
`fix_rho`	A list informing if any rho (rho_s, rho_t, rho_st) is fixed
`range`	Range to use during the MCMC. Entrys are: 'rho_s', 'rho_t' and 'rho_st'
`range`	A list informing a range for sampling rho (rho_s, rho_t, rho_st). Each entre must be a vector in the interval (-1, 1).