Detect similar sounding words in Ruby

Question

I'm aware of SOUNDEX and (double) Metaphone, but these don't let me test for the similarity of words as a whole - for example "Hi" sounds very similar to "Bye", but both of these methods will mark them as completely different.

Are there any libraries in Ruby, or any methods you know of, that are capable of determining the similarity between two words? (Either a boolean is/isn't similar, or numerical 40% similar)

edit: Extra bonus points if there is an easy method to 'drop in' a different dialect or language!

Answer 1

I think you're describing levenshtein distance. And yes, there are gems for that. If you're into pure Ruby go for the text gem.

$ gem install text

The docs have more details, but here's the crux of it:

Text::Levenshtein.distance('test', 'test')    # => 0
Text::Levenshtein.distance('test', 'tent')    # => 1

If you're ok with native extensions...

$ gem install levenshtein

It's usage is similar. It's performance is very good. (It handles ~1000 spelling corrections per minute on my systems.)

If you need to know how similar two words are, use distance over word length.

If you want a simple similarity test, consider something like this:

Untested, but straight forward:

String.module_eval do
   def similar?(other, threshold=2)
    distance = Text::Levenshtein.distance(self, other)
    distance <= threshold
  end
end

Answer 2

You might first preprocess the words using a thesaurus database, which will convert words with similar meaning to the same word. There are various thesaurus databases out there, unfortunately I couldn't find a decent free one for English ( http://www.gutenberg.org/etext/3202 is the one I found, but this doesn't show what relations the specific words have (like similar; opposite; alternate meaning; etc.), so all words on the same line have some relation, but you won't know what that relation is )

But for example for Hungarian there is a good free thesaurus database, but you don't have soundex/metaphone for hungarian texts...

If you have the database writing a program that preprocesses the texts isn't too hard (ultimately it's a simple search-replace, but you might want to preprocess the thesaurus database using simplex or methaphone too)