Arc Forumnew | comments | leaders | submitlogin
2 points by cluelessmoron 5780 days ago | link | parent

      (bayes-train (map lower-case (parse (read-file {big.dms}) {\W} 0)) 'Lexicon)
      (define (edits1 word) 
       (let ((l (length word)) (res '()))
        (for (i 0 (- l 1)) (push (replace (nth i (string word)) (string word) "") res -1))
        (for (i 0 (- l 2)) (push (swap i (+ i 1) (string word)) res -1))
        (for (i 0 (- l 1)) (for (c (char "a") (char "z")) 
          (push (replace (nth i (string word)) (string word) (char c)) res -1)
          (push (replace (nth i (string word)) (string word) (string (char c) (nth i (string word)) )) res -1)))
        res))
      (define (edits2 word) (unique (flat (map edits1 (edits1 word)))))
      (define (known-edits2 word) (filter (fn (w) (Lexicon w)) (edits2 word)))
      (define (known words) (filter (fn (w) (Lexicon w)) words))
      (define (correct word)
        (let ((candidates (or (known (list word)) (known (edits1 word)) (known (edits2 word)))))
          (first (sort candidates (fn (w1 w2) (> (Lexicon w1) (Lexicon w2)))))))
Test

      (join (map correct (parse "tihs sentnce is ful of inkorrect spelingz")) " ")
      
      this sentence is full of incorrect spelling
      
that's newLISP - should be easy to convert...