R/part_matrix_to_vector.R

In jreisner/sparseBiclustering: Biclustering with Missing Data

#' Convert a partition matrix to a vector
#'
#' @param P0 A partition matrix
#' @return An integer vector
#' @description For each row in a partition matrix, this function gets the column index for which the row is equal to one.
#'   That is, for row i, this function returns the index of the row entry that is equal to one.

# get a vector of the row/col prototype each row/col is in (essentially, for
# each row of Q/P, get the column of Q/P that is nonzero (has a 1 in it).)
# pmat2vec.f
part_matrix_to_vector <- function(P0){

  D <- diag(1:ncol(P0))
  PD <- P0 %*% D
  rowSums(PD)

}