SarPoisson, SarNegbin, SarZip, SarZinb:
Poisson, negative binomial (NB2), zero-inflated Poisson and zero-inflated
negative binomial regressions with the spatially lagged mean
\(\exp((I - \rho W)^{-1} X \beta)\), by profile maximum likelihood over
rho (IRLS, dispersion ML and EM for fixed rho). SarPoissonLmTest:
score test of rho = 0. CountModelIc: AIC, BIC, AICc.
BymVarianceFraction: spatial share of the BYM variance with
Sorbye-Rue scaling. Identical to the Python arm morie.fn.spcount.
Usage
SarPoisson(y, X, W, rho_bounds = c(-0.99, 0.99))
SarNegbin(y, X, W, rho_bounds = c(-0.99, 0.99))
SarZip(y, X, W, Z = NULL, rho_bounds = c(-0.99, 0.99))
SarZinb(y, X, W, Z = NULL, rho_bounds = c(-0.99, 0.99))
SarPoissonLmTest(y, X, W)
CountModelIc(loglik, k, n)
BymVarianceFraction(var_spatial, var_unstructured, Q = NULL)Arguments
- y
Counts.
- X
Design matrix (with intercept).
- W
Spatial weights matrix.
- rho_bounds
Search interval for rho.
- Z
Zero-model design (default intercept).
- loglik, k, n
Log-likelihood, number of parameters, sample size.
- var_spatial, var_unstructured
Spatial and unstructured variances.
- Q
ICAR structure matrix (optional).
References
Lambert, D. M., Brown, J. P. and Florax, R. J. G. M. (2010). A two-step estimator for a spatial lag model of counts. Regional Science and Urban Economics 40, 241-252.
Lambert, D. (1992). Zero-inflated Poisson regression. Technometrics 34, 1-14.
Riebler, A., Sorbye, S. H., Simpson, D. and Rue, H. (2016). An intuitive Bayesian spatial model for disease mapping that accounts for scaling. Statistical Methods in Medical Research 25, 1145-1165.