posteriorSummaryOfAUC: Posterior Summary of the AUC of a Multi-scale Occupancy...

In RobertDorazio/eDNAoccupancy: Multi-scale Occupancy Modeling of Environmental DNA Data

Posterior Summary of the AUC of a Multi-scale Occupancy Model's Predictions

Description

Estimates the posterior mean, median, and 95% credible limits for the AUC of a multi-scale occupancy model's predictions.

Usage

posteriorSummaryOfAUC(fit, burnin = 1, mcError = FALSE,
  outputSummary = FALSE)

Arguments

fit	object of class occModel that contains data and previous state of the model's Markov chain
burnin	initial no. iterations of Markov chain to be omitted from calculations
mcError	logical switch to estimate Monte Carlo standard errors
outputSummary	logical switch to return values of posterior summary statistics

Details

This function estimates the posterior mean, median, and 95% credible limits of the AUC (Area Under the Receiver Operating Characteristic curve) of a multi-scale occupancy model's predictions. The AUC of a model is sometimes used in model selection with higher values of AUC favored over lower values.

Value

Prints estimates of summary statistics for the posterior distribution of the AUC of a multi-scale occupancy model's predictions. If mcError=TRUE, the Monte Carlo standard errors of these estimates are computed. If outputSummary=TRUE, the posterior summaries are returned in a list.

Examples

data(gobyDetectionData)
detections = occData(gobyDetectionData, "site", "sample")
data(gobySurveyData)
gobySurveyData = scaleData(gobySurveyData)  # center and scale numeric covariates

fit1 = occModel(formulaSite          = ~ veg,
                formulaSiteAndSample = ~ sal + twg,
                formulaReplicate     = ~ sal + fish,
                detectionMats        = detections,
                siteData             = gobySurveyData,
                niter                = 1100,
                niterInterval        = 100,
                siteColName = 'site',
                )

posteriorSummaryOfAUC(fit1, burnin=100, mcError=TRUE)

RobertDorazio/eDNAoccupancy documentation built on Sept. 5, 2023, 9:57 a.m.