Skip to contents

Conduct goodness-of-fit test for spOccupancy models

diagnoseGofSpOccu.Rd

This function comes largely from the ppcOcc.R from the spOccupancy package

Usage

diagnoseGofSpOccu(object, post_sims, fit_stat = "chi-squared", group = 1)

Arguments

object

an spOccupancy fit

post_sims

A list of posterior simulations obtained with simDetSpOccu().

fit_stat

Goodness of fit statistic to compute. Currently only Chi-squared is supported.

group

Whether the GOF statistic should be computed for sites (1) or for visits (2). Currently, only grouping by site is supported

Value

A list GOF statistics computed for observed data and for simulated data. One statistic for is computed for each MCMC iteration. The function will create a list with several objects. The most important are:

  • fit.y: For each MCMC sample, chi-squared statistic for total number of detections for all sites in data wrt expected model expectation

  • fit.y.rep: For each MCMC sample, chi-squared statistic for total number of detections for all sites in simulated data wrt expected model expectation

  • fit.y.group.quants: For each site, chi-squared statistic for total number of detections in data wrt expected model expectation (posterior distribution quantiles)

  • fit.y.rep.group.quants: For each site, chi-squared statistic for total number of detections in simulated data wrt expected model expectation (posterior distribution quantiles)

  • y.summ.per.site: Data frame with number of detections per site obtained from data simulated from the posterior distribution. Also the observed number of detections in the data.