README.md

In tmcd82070/evoab: Evidence of Absence (EoA)

Evidence of Absence Regression Package

These are routines implement Evidence of Absence Regression (EoAR) which estimates the number of missed entities after a series of searches and relates them to covariates. Special cases are the Evidence of Absence (EoA) model of Huso et al. (2015) and the Informed Evidence of Absence (IEoA) approaches.

How to git it:

From GitHub, issue the following commands in a git shell:

cd <directory you want> git clone https://github.com/tmcd82070/evoab.git

The above commands will download all source from GitHub to your computer. Among other things, you should see a DESCRIPTION file and R directory.

Intalling:

Using devtools

Open R and setwd() to the directory containing the DESCRIPTION file. In R issure the following:

library(devtools) document() install()

Manual install

Open a command window, change directory to the folder containing DESCRIPTION and issue the following command:

r CMD INSTALL evoab

To Contribute

If you change something, and it's useful, I would be very interested in hearing about it.

Usage Example

The main routine is eoa. It takes a count vector, model for lambda, and g-values, Here is an example :

This is fake data from a three year study on seven sites. First, the alpha and beta parameters for g-value distributions, one per year. ns <- 3 ny <- 7 g <- data.frame( alpha = rnorm(ns*ny,70,2), beta = rnorm(ns*ny,700,25) )

This is the carcasses count vector, one count per site per year.

Y <- rbinom(ns*ny, c(rep(20,ny), rep(40,ny), rep(60,ny)), g$alpha/(g$alpha+g$beta))

This is the covariate data frame. This data frame contains Year as a linear effect (1,2,3,etc) and Year as a factor (2015, 2016, 2017, etc).

df <- data.frame(year=factor(c(rep("2015",ny),rep("2016",ny),rep("2017",ny))), Year=c(rep(1,ny),rep(2,ny),rep(3,ny)))

The following relates carcass deposition rates to year using vague priors for coefficients:

eoa.1 <- eoa(Y~year, g, df )

The following uses informed distributions:

# Assume prior mean is 10 and prior sd is 3 # Fit intercept-only model to get one mean lambda intMean <- 2*log(10) - 0.5*log(3^2 + 10^2) intSd <- sqrt(-2*log(10) + log(3^2 + 10^2)) prior <- data.frame(mean=intMean, sd=intSd, row.names="(Intercept)") eoa.1 <- eoa(Y~1, g, df, priors=prior )

After either run, you should check convergence. To do so, run a traceplot and Gelman stats. Any Rhats > 1.1 indicate suspect convergence. The following commands are useful for inspecting mixing and convergence:

library(lattice) xyplot(ieoa.1$out[,labels(ieoa.1)]) acfplot(ieoa.1$out[,labels(ieoa.1)]) densityplot(ieoa.1$out[,labels(ieoa.1)]) gelman.diag(ieoa.1$out) # gelmanStats gelman.plot(ieoa.1$out) # gelmanPlot

tmcd82070/evoab documentation built on May 13, 2020, 11:25 p.m.