R/rCensspatial_USER.R

In RcppCensSpatial: Spatial Estimation and Prediction for Censored/Missing Responses

#' Censored spatial data simulation
#'
#' It simulates censored spatial data with a linear structure for an established censoring rate.
#'
#' @param beta linear regression parameters.
#' @param sigma2 partial sill parameter.
#' @param phi spatial scaling parameter.
#' @param nugget nugget effect parameter.
#' @param x design matrix of dimensions \eqn{n\times q}.
#' @param coords 2D spatial coordinates of dimensions \eqn{n\times 2}.
#' @param cens \code{'left'} or \code{'right'} censoring. By default \code{='left'}.
#' @param pcens desired censoring rate. By default \code{=0.10}.
#' @param npred number of simulated data used for cross-validation (Prediction). By default \code{=0}.
#' @param cov.model type of spatial correlation function: \code{'exponential'}, \code{'gaussian'},
#' \code{'matern'}, and \code{'pow.exp'} for exponential, gaussian, matérn, and power exponential, respectively.
#' @param kappa parameter for some spatial correlation functions. For exponential and
#' gaussian \code{kappa=NULL}, for power exponential \code{0 < kappa <= 2}, and for
#' matérn correlation function \code{kappa > 0}.
#'
#' @return If \code{npred > 0}, it returns two lists: \code{Data} and
#' \code{TestData}; otherwise, it returns a list with the simulated data.
#'
#' \code{Data}
#' \item{y}{response vector.}
#' \item{ci}{censoring indicator.}
#' \item{lcl}{lower censoring bound.}
#' \item{ucl}{upper censoring bound.}
#' \item{coords}{coordinates matrix.}
#' \item{x}{design matrix.}
#'
#' \code{TestData}
#' \item{y}{response vector.}
#' \item{coords}{coordinates matrix.}
#' \item{x}{design matrix.}
#'
#' @author Katherine L. Valeriano, Alejandro Ordoñez, Christian E. Galarza, and Larissa A. Matos.
#'
#' @examples
#' n = 100
#' set.seed(1000)
#' coords = round(matrix(runif(2*n,0,15),n,2), 5)
#' x = cbind(1, rnorm(n))
#' data = rCensSp(beta=c(5,2), sigma2=2, phi=4, nugget=0.70, x=x,
#'                coords=coords, cens="left", pcens=0.10, npred=10,
#'                cov.model="gaussian")
#' data$Data
#' data$TestData

rCensSp = function(beta, sigma2, phi, nugget, x, coords, cens="left", pcens=0.10, npred=0,
                   cov.model="exponential", kappa=NULL){
  beta = c(beta)
  if (!is.numeric(beta)) stop("beta must be a numeric vector")
  if (!all(is.finite(beta))) stop("beta must contain only finite values")
  if (length(c(sigma2))>1 | !is.numeric(sigma2)) stop("sigma2 must be a non-negative number")
  if (sigma2 <= 0) stop("sigma2 must be non-negative")
  if (length(c(phi))>1 | !is.numeric(phi)) stop("phi must be a non-negative number")
  if (phi <= 0) stop("phi must be non-negtaive")
  if (length(c(nugget))>1 | !is.numeric(nugget)) stop("nugget must be a non-negative number")
  if (nugget <= 0) stop("nugget must be non-negative")

  x = as.matrix(x)
  if (!all(c(is.finite(x)))) stop("x must contain only finite values")
  if (ncol(x) != length(c(beta))) stop("Non-conformable dimensions between x and beta")

  coords = as.matrix(coords)
  if (!all(c(is.finite(coords)))) stop ("coords must contain only finite values")
  if (ncol(coords) != 2) stop("coords must contain 2 columns")
  if (nrow(coords) != nrow(x)) stop("Non-conformable dimensions between coords and x")

  if (is.null(cens)) stop("cens must be specified")
  if (cens!="left" & cens!="right") stop("cens should be one of left or right")

  if (!is.numeric(pcens) | length(c(pcens))>1) stop ("pcens must be a real number in [0,1]")
  if (pcens<0 | pcens>1) stop("pcens must be a real number in [0,1]")

  if (!is.numeric(npred) | length(c(npred))>1) stop ("npred must be a positive integer")
  if (npred<0 | npred>=nrow(x) | npred%%1!=0) stop("npred must be a positive integer")

  if (is.null(cov.model)) stop("cov.model must be specified")
  if (cov.model!="matern" & cov.model!="gaussian" & cov.model!="pow.exp" & cov.model!="exponential"){
    stop("cov.model should be one of matern, gaussian, pow.exp or exponential")}

  if (cov.model!="exponential" & cov.model!="gaussian"){
    if (length(c(kappa))>1 | !is.numeric(kappa)) stop("kappa must be specified")
    if (cov.model=="pow.exp" & (kappa>2| kappa<=0)) stop("kappa must be a real in (0,2]")
    if (cov.model=="matern" & kappa<=0) stop("kappa must be a real number in (0,Inf)")
    if (cov.model=="matern" & is.infinite(kappa)) stop("kappa must be a real number in (0,Inf)")
  } else { kappa=0 }

  out.gen = random.cens(beta,sigma2,phi,nugget,x,coords,cens,pcens,npred,cov.model,kappa)
  return(out.gen)
}