R/rouge_n.R

In LJCovingtonJr/rougeR: A Package for Automatic Evaluation of Summaries

#' Rouge-N: N-gram Co-occurrence Statistics
#'
#' Takes a candidate summary, reference summary, and n value, returning the associated ROUGE-N recall, precision, and F values.
#' @param candidate the candidate (usually machine generated) summary
#' @param reference the reference (usually human written) summary
#' @param n the number of grams to be considered, default 1.
#' @return a dataframe, consisting of a single observation and three variables: recall, precision, and f_measure.
#' @import tokenizers
#' @export

rouge_n <- function(candidate, reference, n = 1){
  candidate_ngrams <- unlist(tokenizers::tokenize_ngrams(candidate, n=n))
  reference_ngrams <- unlist(tokenizers::tokenize_ngrams(reference, n=n))

  overlap <- sum(candidate_ngrams %in% reference_ngrams)

  recall <- (overlap/length(reference_ngrams))
  precision <- (overlap/length(candidate_ngrams))
  f_measure <- f_measure(recall, precision)

  result <- data.frame(recall, precision, f_measure)
  return(result)
}