expsim_update_parameter: Update Exposure Model Parameters

In jpkeller/bercs: Bayesian Exposure-Response Curves via Stan

Description

Sets or overwrites data-generating parameters for an exposure model object

Usage

expsim_update_parameter(
  obj,
  param,
  level = c("group", "cluster", "unit", "observation", "correlation", "time"),
  type = c("mean", "sd", "re"),
  value = NULL,
  draw = is.null(value)
)

Arguments

obj	'expsim' object to update
param	character string giving name of parameter to update. If not provided, parameter name is taken from level and type
level	character string of model level being updated. Must be one of "group", "cluster", "unit", "observation", or "time". See Details.
type	character string giving type of parameter being updated. See Details.
value	value(s) to which parameter should be set
draw	if TRUE, then the parameter value is sampled instead of being set to value

Details

Only specific combinations of level and type are allowed. They are:

• "group": "mean" sets 'etaG'

• "cluster": "sd" and "re" set 'sigK' and 'reK', respectively

• "unit": "sd" and "re" set 'sigI' and 'reI', respectively

• "observation": "sd" sets 'sigW'.

• "time": "mean" sets 'timefn'.

See Also

Other exposure simulation functions: create_exposure_simulation_skeleton(), expsim_sample_observations(), sim_update_times()

Examples

# Create simulation structure
es <- create_exposure_simulation_skeleton_parallel(ngroups=2, nunits=10)

# No standard deviation parameter set by default
es$structure$sigK
# Add standard deviation parameter value
es <- expsim_update_parameter(es, level="cluster", type="sd", value=1)
es$structure$sigK
# Sample random effect values
es <- expsim_update_parameter(es, level="cluster", type="re", draw=TRUE)

# Add time function
es <- expsim_update_parameter(es, level="time", type="mean", value=function(w) 2*cos(w))

jpkeller/bercs documentation built on March 24, 2021, 5:36 a.m.