R/predict-msaenet.R

In road2stat/msaenet: Multi-Step Adaptive Estimation Methods for Sparse Regressions

#' Make Predictions from an msaenet Model
#'
#' Make predictions on new data by a msaenet model object.
#'
#' @param object An object of class \code{msaenet} produced
#' by \code{\link{aenet}}, \code{amnet}, \code{asnet},
#' \code{\link{msaenet}}, \code{\link{msamnet}}, or \code{\link{msasnet}}.
#' @param newx New data to predict with.
#' @param ... Additional parameters, particularly prediction \code{type} in
#' \code{\link[glmnet]{predict.glmnet}}, \code{\link[ncvreg]{predict.ncvreg}},
#' or \code{\link[ncvreg]{predict.ncvsurv}}.
#'
#' @return Numeric matrix of the predicted values.
#'
#' @method predict msaenet
#'
#' @author Nan Xiao <\url{https://nanx.me}>
#'
#' @importFrom stats predict
#'
#' @export
#'
#' @examples
#' dat <- msaenet.sim.gaussian(
#'   n = 150, p = 500, rho = 0.6,
#'   coef = rep(1, 5), snr = 2, p.train = 0.7,
#'   seed = 1001
#' )
#'
#' msaenet.fit <- msaenet(
#'   dat$x.tr, dat$y.tr,
#'   alphas = seq(0.2, 0.8, 0.2),
#'   nsteps = 3L, seed = 1003
#' )
#'
#' msaenet.pred <- predict(msaenet.fit, dat$x.te)
#' msaenet.rmse(dat$y.te, msaenet.pred)
predict.msaenet <- function(object, newx, ...) {
  if (missing(newx)) stop("Please specify `newx` to predict on.", call. = FALSE)

  if (!.is.msaenet(object)) stop(message.object.type, call. = FALSE)

  if (.is.glmnet(object$"model")) {
    pred <- predict(object$"model", newx = newx, ...)
  }

  if (.is.ncvreg(object$"model")) {
    pred <- predict(object$"model", X = newx, ...)
  }

  pred
}