pcvreg function

Procrustes cross-validation for multivariate regression models

This is a generic method, use pcvpls() or pcvpcr() instead.

pcvreg(
  X,
  Y,
  ncomp = min(nrow(X) - 1, ncol(X), 30),
  cv = list("ven", 4),
  center = TRUE,
  scale = FALSE,
  funlist = list(),
  cv.scope = "global"
)

Arguments

• X: matrix with predictors from the training set.
• Y: vector with response values from the training set.
• ncomp: number of components to use (more than the expected optimal number).
• cv: which split method to use for cross-validation (see description of method pcvpls() for details).
• center: logical, to center or not the data sets
• scale: logical, to scale or not the data sets
• funlist: list with functions for particular implementation
• cv.scope: scope for center/scale operations inside CV loop: 'global' — using globally computed mean and std or 'local' — recompute new for each local calibration set.