fit.jagsMANEC: fit.jagsMANEC

In AIMS/NEC-estimation: A Bayesian No Effect Concentration (NEC) package. R package version 1. doi:10.5281/ZENODO.3966864

Description

Fits a variety of NEC models using jags and provides a model averaged predictions based on DIC model weights

Usage

fit.jagsMANEC(
  data,
  x.var,
  y.var,
  trials.var = NA,
  x.type = NA,
  y.type = NA,
  burnin = 5000,
  n.iter = burnin + 500,
  n.iter.update = 10000,
  n.tries = 3,
  params = c("top", "beta", "NEC", "SS", "SSsim"),
  over.disp = FALSE,
  sig.val = 0.01,
  model.set = "all",
  ...
)

Arguments

data	a data.frame containing the data to use for the model
x.var	the column heading indicating the concentration (x) variable
y.var	the column heading indicating the response (y) variable
trials.var	the column heading indicating the column for the number of "trials" for binomial response data. If not supplied, the model may run but will not be the model you intended!
x.type	the statistical distribution to use for the x (concentration) data. This will be guess based on the characteristic of the input data if not supplied.
y.type	the statistical distribution to use for the y (response) data. This may currently be one of 'binomial', 'poisson',' 'gaussian', 'beta', 'negbin' or 'gamma'. Others can be added as required, please contact the package maintainer. If not supplied, the appropriate distribution will be guessed based on the distribution of the input data.
burnin	the number of iterations to use as burnin.
n.iter	the number of iterations to run following burnin for the initial jags fit. Defaults to 500 + burnin,
n.tries	the number of tries to attempt to fit the model and attain good chain mixing. See details below.
params	a vector of names indicating the parameters to trace during the jags fit.
sig.val	probability value to use as the lower quantile to test significance of the predictor posterior values against the control, to estimate NSEC as an interpolated NOEC value from smooth ECx curves.
model.set	A vector of the names of model types to be fit. Currently defaults to all available model types. If "NEC" is supplied, only the NEC models will be fit. If "ECx" is supplied, only continuous curve models will be fit. see ?fit.jagsNEC to see the available models to fit.

Value

All successully fitted jags model fits, mod.stats a data.frame of model fit statistics, NEC a model averaged posterior of the estimated NEC, and pred.vals a list of model averaged predictions.

AIMS/NEC-estimation documentation built on Dec. 7, 2020, 10:44 a.m.