R/CVMSE.R

In CorReg: Linear Regression Based on Linear Structure Between Variables

#' Cross validation
#' 
#' @param X covariates matrix (double)
#' @param Y response variable
#' @param K number of classes
#' @param intercept (boolean) with or without an intercept
#' @param methode the methode used by OLS.
#' @param groupe a vector to define the groups used for cross-validation (to obtain a reproductible result)
#' 
#' @export
CVMSE <- function(X = X, Y = Y, K = K, intercept = TRUE, methode = 1, groupe = NULL) {
  K = abs(K)
  K = min(K, nrow(X))
  Y = as.double(Y)
  if (is.null(groupe)) {
    groupe = rep(0:(K - 1), length.out = nrow(as.matrix(X)))
    groupe = sample(groupe)
  }
  res = .Call("CVMSE", X, Y, K, intercept, methode, groupe, PACKAGE = "CorReg")
  return(res$MSE)
}