create_standata_exposure: Create List for Fitting Exposure Model via STAN

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

Description

Creates a list object that follows the structure required by the exposure models that are implemented in STAN.

Usage

create_standata_exposure(
  data = NULL,
  group = data$group,
  conc = data$conc,
  unit_id = data$unit_id,
  clust_id = data$clust_id,
  time = data$time,
  Mt = NULL,
  log_transform = FALSE,
  return_addition = FALSE
)

Arguments

data	Optional data frame containing data in a long format.
group	Vector of group assignments. See Details.
conc	Vector of exposure concentrations.
unit_id	Vector identifying the distinct unit (e.g. person, household) for each observation.
clust_id	Vector identifying cluster membership for each observation.
time	Vector of times corresponding to conc values.
Mt	Matrix of splines values for time to include in the model. Defaults to NULL, and can be added later via add_spline_time. See create_spline for creating this matrix.
log_transform	Should the concentration value be log-transformed?
return_addition	See 'Value'.

Details

This function takes as input a 'long' data frame and extracts from it the information needed for fitting the exposure model in STAN.

The values of group, unit_id, and clust_id are converted to factor and then coerced to integer. The ordering of groups (units, clusters) in the output object will depend on the default ordering introduced by factor.

Value

A list of class standata_exposure that contains the following:

• G – Number of groups

• K – Number of clusters.

• n Number of households

• N – Number of observations

• cluster_of_obs – Integer providing cluster number of each observation

• group_of_obs – Integer providing group number of each observation

• hh_of_obs – Integer providing household number of each observation

• w – Concentration observations

• times – The date/time of each observation. This is not used directly in model fitting, but contained in the object to facilitate plotting and other summaries.

If return_addition=TRUE, then a two-element list is returned. The first element is the standata_exposure object described above. The second element is a modified version of data, with the variables group_of_obs, cluster_of_obs, hh_of_obs, and times added (or overwritten).

See Also

sample_exposure_model, create_spline, add_spline_time, create_standata_outcome

Examples

# Create simulated data
exp_data <- create_standata_exposure(group=rep(1, 10),
                                     conc=rnorm(10),
                                     unit_id=rep(0:1, 5),
                                     time=runif(10))
# Add default priors
exp_data <- add_priors(exp_data,
                       sigmaI=c(0, 0.1))
# Fit the model
exp_mod_fit <- sample_exposure_model(exp_data)

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