bayesnec (version 2.1.3.0)

update.bnecfit: Update an object of class bnecfit as fitted by function bnec.

Description

Update an object of class bnecfit as fitted by function bnec.

Usage

# S3 method for bnecfit
update(
  object,
  newdata = NULL,
  recompile = NULL,
  x_range = NA,
  resolution = 1000,
  sig_val = 0.01,
  loo_controls,
  force_fit = FALSE,
  ...
)

Value

An object of class bnecfit. If one single model is returned, then also an object of class bayesnecfit; otherwise, if multiple models are returned, also an object of class bayesmanecfit.

Arguments

object

An object of class bnecfit as fitted by function bnec.

newdata

Optional data.frame to update the model with new data. Data-dependent default priors will not be updated automatically.

recompile

A logical, indicating whether the Stan model should be recompiled. If NULL (the default), update tries to figure out internally, if recompilation is necessary. Setting it to FALSE will cause all Stan code changing arguments to be ignored.

x_range

A range of predictor values over which to consider extracting ECx.

resolution

The length of the predictor vector used for posterior predictions, and over which to extract ECx values. Large values will be slower but more precise.

sig_val

Probability value to use as the lower quantile to test significance of the predicted posterior values against the lowest observed concentration (assumed to be the control), to estimate NEC as an interpolated NOEC value from smooth ECx curves.

loo_controls

A named list of two elements ("fitting" and/or "weights"), each being a named list containing the desired arguments to be passed on to loo (via "fitting") or to loo_model_weights (via "weights"). If "weights" is not provided by the user, bnec will set the default method argument in loo_model_weights to "pseudobma". See ?loo_model_weights for further info.

force_fit

Should model truly be updated in case either newdata of a new family is provided?

...

Further arguments to brm.

Examples

if (FALSE) {
library(bayesnec)
data(manec_example)
# due to package size issues, `manec_example` does not contain original
# stanfit DSO, so need to recompile here
smaller_manec <- update(manec_example, chains = 2, iter = 50,
                        recompile = TRUE)
# original `manec_example` is fit with a Gaussian
# change to Beta distribution by adding newdata with original `nec_data$y`
# function will throw informative message.
beta_manec <- update(manec_example, newdata = nec_data, recompile = TRUE,
                     chains = 2, iter = 50,
                     family = Beta(link = "identity"), force_fit = TRUE)
}