Computes the value of WAIC (Widely Applicable Information Criterion) for the Hmsc model
computeWAIC(hM, ghN =11, byColumn =FALSE)
Arguments
hM: a fitted Hmsc model object
ghN: order of Gauss-Hermite quadrature for approximate numerical integration
byColumn: describes whether WAIC is computed for the entire model byColumn=FALSE or for each column (i.e. species) byColumn=TRUE
Returns
the scalar WAIC
Details
The result is exact for normal and probit observational models. For Poisson-type observational model the result is obtained through numerical integration using Gauss-Hermite quadrature.
Examples
# Compute WAIC of previously sampled Hmsc objectWAIC = computeWAIC(TD$m)