R/pre_process_data.R

In jinhyunju/icreport: Tools for creating a visual summary of matrix decomposition results.

#' Pre-processing of the data
#'
#' The function takes in a g x N matrix, where g is the number of features and N is the number of samples,
#' and removes any 0 variance entries followed by mean centering.
#' If scaling option is true it will scale all rows to have unit standard deviation.
#'
#' @param input_matrix Input matrix with a dimension of g X N  \code{input_matrix}
#' @param scale_pheno If TRUE features are scaled to have unit standard deviation.
#' @return Pre-processed data matrix
#' @keywords keywords
#'
#' @examples
#' raw.data <- rbind(rnorm(100,2,3),rnorm(100,10,2))
#' centered <- pre_process_data(raw.data, scale_pheno = FALSE)
#' apply(centered, 1, mean)
#' apply(centered, 1, sd)
#' centered.scaled <- pre_process_data(raw.data, scale_pheno = TRUE)
#' apply(centered.scaled, 1, mean)
#' apply(centered.scaled, 1, sd)
#' @export
pre_process_data <- function(input_matrix, scale.pheno){
    var.vec <- apply(input_matrix, 1, var)
    var0.count <- sum(var.vec == 0)

    # Remove 0 variance  if existing
    if(var0.count != 0){
        message("- Removing ",var0.count," 0 variance features \n")
        input_matrix <- input_matrix[-c(which(var.vec == 0)),]
    } else {
        message("- No 0 variance features in dataset \n")
    }


    if(scale.pheno == TRUE){
        message("- Centering and Scaling Input Matrix \n")
        # scale Input matrix to have 0 mean and 1 sd
        input_matrix <- t(apply(input_matrix, 1, function(x) (x - mean(x))/sd(x)))
    } else {
        message("- Centering Input Matrix without Scaling \n")
        input_matrix <- t(apply(input_matrix, 1, function(x) (x - mean(x))))
    }

    message("- Dimension of Input Matrix = [",
            nrow(input_matrix),",",ncol(input_matrix),"]\n")
    return(input_matrix)
}