model_setup: Set up a casecrossover model for fitting.

In awstringer1/casecrossover: Approximate Bayesian Inference for Case Crossover Models

Description

Set up a casecrossover model for fitting: creates all internal objects and stores them in a list with class "cc_model_data". May contain standard linear terms "x" parsable by model.matrix() and smooth terms "s(x)" to which RW2 models are fit (see documentation for s()).

The data will be sorted using dplyr::arrange(id,case) where id and case are the names of the variables identifying the subject and the case day respectively.

The case column in your dataframe should contain a 0 for control days, and a positive integer for case days. If the case day is a 1, then the model is fit exactly as described in the paper; if it is a positive integer greater than 1, then this is equivalent to having a dataset where that subject is replicated that many times (i.e. the full log-likelihood for that subject is multiplied by case). This is a convenience to save dataset storage space/memory.

Usage

model_setup(formula, data, control = cc_control(), verbose = FALSE)

Arguments

formula	A standard R formula; see description. Must contain a strata() term which defines the variable in "data" used to distringuish individual subjects.
data	A data.frame or tibble containing at minimum one column which groups subjects together, one column indicating the case day, and one or more columns containing covariates.
control	A list containing control parameters. See cc_control().
verbose	Logical. Print progress and diagnostic information? Useful for debugging or keeping up with what the function is doing.

Value

An object of class cc_modeldata, for passing into future functions.

Examples

model_setup(y~x + strata(id),data.frame(y = c(0,1),x = c(1,2),id = c(1,1)))

awstringer1/casecrossover documentation built on March 11, 2021, 4:41 a.m.