R/get_split_sets.R

In sRDA: Sparse Redundancy Analysis

get_split_sets <- function(X, Y, nr_subsets){
#Function to split up data matrix into n subsets
#
#input:     X             n*p matrix      - data matrix to be split
#           Y             n*q matrix      - data matrix to be split
#           nr_subsets    integer         - nr of subsets the data
#                                           matrix will be split to
#
#
#output:    label         vector with n_subset unique elements with length of n
#
#           X.sampled,    n*p matrix      shuffled dataset of the origninal data
#           Y.sampled,    n*q matrix      shuffled dataset of the origninal data


  n <- dim(X)[1]

  #re-sample data rows
  splitting_dimensions <- sample(1:n,n)

  X.sampled <- X[splitting_dimensions,]
  Y.sampled <- Y[splitting_dimensions,]

  #calculate how many rows are left over
  leftover = n %% nr_subsets

  rep_eat = (n-leftover)/nr_subsets

  #repeat sequence nr_subsets
  labels = rep(1:nr_subsets , each=rep_eat)

  if(leftover!=0)
  {
    labels = c(labels, (1:nr_subsets)[1:leftover])
  }

  #**********************
  result <- list(X.sampled = X.sampled,
                 Y.sampled = Y.sampled,
                 labels = labels
  )


  result

}