R/AllClass.R

In Biostatistics4SocialImpact/dlm: Distributed Lag Models for Built Environment Applications

## LagBasis
## -------------------------------------------------------------------
#' @title Create and manipulate lag basis functions
#' @inherit basis references
#'
#' @description
#' S4 class object to store and query components of computed
#' lag basis functions. User interface for creating this class can be found in
#' \code{\link{basis}}.
#'
#' @slot x
#'   radii that define ring-shaped areas around participant locations
#' @slot x.center
#'   the value \code{x} was centered to prior to computing the basis
#' @slot x.scale
#'   the value \code{x} was scaled by prior to computing the basis
#' @slot C0
#'   \eqn{C_0} part of basis matrix
#' @slot K1
#'   \eqn{K_1} part of basis matrix
#' @slot dist.fun
#'   the function applied to the elements of \code{x} to compute the basis
#'
#' @details
#' See \code{\link{basis}} for details of the decomposition
#'
#' @name LagBasis
#'
LagBasis <- setClass("LagBasis",
         slots = c(
           x  = "numeric",
           x.center = "numeric",
           x.scale = "numeric",
           C0 = "matrix",
           K1 = "matrix",
           dist.fun = "function"
         ),
         prototype = list(
           x = NA_real_,
           x.center = 0,
           x.scale = 1,
           C0 = matrix(),
           K1 = matrix()
         ))




## SmoothLag
## -------------------------------------------------------------------
#' @title Lag matrix with applied smoothing
#'
#' @description
#' An S4 object for representing lag covariates and
#' storing details about the basis smoothing. Intended for use within the
#' \code{\link{dlm}} modeling framework to assist extraction of basis
#' components treated as "fixed" and "random" effects. Inherits from
#' \code{matrix}.
#'
#' @slot basis
#'   A \code{\link{LagBasis}} smoothing object containing details about the
#'   lag and the smoothing parameters used
#' @slot .Data
#'   Contains the "fixed effects" components of the smoothed lag function.
#'   This scheme is intended to work conveniently with
#'   \code{stats::\link[stats]{model.matrix}}
#' @slot random
#'   Contains the random effects or penalized components of the smoothed
#'   lag function
#' @slot signature
#'   Character vector containing the function name and deparsed arguments
#'   from whatever smoothing function generated the \code{SmoothLag} object
#'
#' @name SmoothLag
#'
SmoothLag <- setClass("SmoothLag",
                       slots = c(
                         basis = "LagBasis",
                         random = "matrix",
                         signature = "character"
                       ),
                       contains = "matrix"
                       )




## dlMod
## -------------------------------------------------------------------
#' @title Distributed lag model objects
#'
#' @description
#' A fitted distributed lag model object. Inherits from \pkg{lme4}'s
#' \code{\link[lme4]{merMod}} so that most methods defined for this
#' parent class should work seamlessly within \pkg{dlmBE} analysis
#'
#' @slot resp
#'   An \code{lme4::\link[lme4]{lmResp}} object to store a
#'   (mixed) model response variable
#' @slot bases
#'   A list of \code{\link{LagBasis}} objects corresponding
#'   to the unique set of lag bases used to fit the model
#' @slot index
#'   A (named) integer vector providing the index of the
#'   basis set in \code{bases} corresponding to each lag term in the
#'   model
#'
#' @name dlMod
#'
dlMod <- setClass("dlMod",
                  slots = c(
                    resp = "lmResp",   # from lme4
                    bases = "list",
                    index = "numeric"
                  ),
                  contains = "merMod"  # also from lme4
                  )