R/rdist.R

In thiyangt/fformpp: Feature-based FORecast Model Performance Prediction

##' Calculate the distance between all rows of X and Y.
##' @title Euclidean distance matrix between two matrices
##' @name  rdist
##' @param X an `m-by-p` numeric matrix.
##' @param Y an `n-by-p` numeric matrix.
##' @param log logical; If TRUE, the log distance matrix is returned.
##' @return an `m-by-n` numeric matrix.
##' @export
rdist <- function(X, Y, log = TRUE)
{
  dim.X <- dim(X)
  dim.Y <- dim(Y)
  if(dim.X[2] != dim.Y[2])
  {
    stop("Input X and Y dimension does not match!")
  }
  sum.X <- matrix(rowSums(X^2), dim.X[1], dim.Y[1])
  sum.Y <- matrix(rowSums(Y^2), dim.X[1], dim.Y[1], byrow = TRUE)
  dist0 <- sum.X + sum.Y - 2*tcrossprod(X, Y)
  dist0[dist0<0] <- 0 # error handling for numerical inconsistent.

  if(log == TRUE) ## return log distance
  {
    out <- 1/2*log(dist0)
  }
  else
  {
    out <- sqrt(dist0)
  }

  return(out)
}