R/Interpretor.R

In CorReg: Linear Regression Based on Linear Structure Between Variables

# 'another way to interprete the structure
Interpretor <- function(Z = Z, names = NULL, X = NULL, R2 = TRUE, R2_vect = NULL, adj = TRUE)
{
  if (R2) {
    if (is.null(R2_vect)) {
      R2_vect = R2Z(Z = Z, X = X, adj = adj, crit = "R2")
    }
    ordre = order(R2_vect)
    X = X[, ordre]
    Z = Z[ordre, ordre]
    names = names[ordre]
  }

  if (is.null(names)) {
    if (!is.null(X)) {
      names = names(X)
    } else {
      message("no names")
      # break
    }
  }

  groups = list()
  quiI2 = which(colSums(Z) != 0)
  compt = 1
  for (j in quiI2) {
    groups[[compt]] = c(names[j], names[which(Z[, j] != 0)])
    compt = compt + 1
  }
  if (R2) {
    R2_vect = R2_vect[ordre]
    names(groups) = paste(1:length(groups), R2_vect[R2_vect != 0])
  }

  return(groups)
}