Example code for using recursive Bayesian computation in optimal design, using simulated data and binary regression.

Leach, Clinton B., Perry J. Williams, Joseph M. Eisaguirre, Jamie N. Womble, Michael R. Bower, and Mevin B. Hooten, submitted 2021, Recursive Bayesian computation facilitates adaptive optimal design in ecological studies, Ecology

output/data.rds
scripts/plots.R
src/designs.jl
src/example.jl
src/probit.reg.jl
optimal-design.Rproj
LICENSE
README.md

output/data.rds – R data file containing a data frame of simulated data with columns loc – integer site index; y – binary presence/absence at the site; x – covariate value at the site; K1 – binary indicator of inclusion in the initial sample; score – computed design criteria for including that site in a second data collection effort

scripts/plots.R – R script for generating Figure 1 of the above manuscript

src/designs.jl – Julia code defining a function to loop over designs and predicted data, apply PPRB, and compute quantities needed to evaluate the design criteria

src/example.jl – Julia script simulating binary data, fitting an initial binary regression model, and computing the design criteria for each of the potential designs

src/probi.reg.jl – Julia code defining functions to fit an initial probit regression model to a complete data set (probit_reg) and to use PPRB to update an initial model fit given new data (probit_pprb)

optimal-design.Rproj – Rproject file for loading the project in Rstudio

LICENSE – file specifying the license of the included code

README.md – file containing same metadata