Find number of matching words in phrases.

gistfile1.r

# match on intersecting word set
matching_words = function(x, y) {

  x = unlist(strsplit(x, ' ', fixed=TRUE))
  y = strsplit(y, ' ', fixed=TRUE)

  # get intersections
  results = sapply(y, function(b) { length(intersect(x, b)) })

  # break ties in intersection counts based on simplicity (length) of original phrase
  results_freq = table(results)
  results_freq = results_freq[names(results_freq) > 0]

  y_length = sapply(y, length)

  for (f in names(results_freq)) {
    mask = which(results == f)
    results[mask] = 1.0 - (rank(y_length[mask]) - 1.0) / results_freq[f]
  }

  return(results)
}