Posterior dispersion — dispersion • bayesnec

Calculates a posterior dispersion metric.

dispersion(model, summary = FALSE, seed = 10)

Arguments

model: An object of class bayesnecfit whose distribution family is either poisson or binomial.
summary: Logical. Should summary stats be returned instead of full vector? Defaults to FALSE.
seed: Change seed for reproducible purposes.

Value

A numeric vector. If summary is FALSE, an n-long vector containing the dispersion metric, where n is the number of post warm-up posterior draws from the brmsfit object. If TRUE, then a data.frame containing the summary stats (mean, median, 95% highest density intervals) of the dispersion metric.

Details

This function calculates a dispersion metric which takes the ratio between the observed relative to simulated Pearson residuals sums of squares.

References

Zuur, A. F., Hilbe, J. M., & Ieno, E. N. (2013). A Beginner's Guide to GLM and GLMM with R: A Frequentist and Bayesian Perspective for Ecologists. Highland Statistics Limited.

Examples

if (FALSE) {
library(bayesnec)
data(nec_data)
nec_data$y <- as.integer(round(nec_data$y * 100))
nec4param <- bnec(y ~ crf(x, "nec4param"), data = nec_data, chains = 2)
dispersion(nec4param, summary = TRUE)
}