Match Algorithm: How close of a match

[zer]

Good morning,

Is there anything available that I can compair two strings to determine how closely they match each other? I am thinking there should be a perl module for this but I can't seem to find any.

Alternately, I would likely program this myself. Perhaps there is an algorithm or some experience in the monestary that could point me in the right direction so I don't end up recreating the wheel.

I have text that was munged up by an application. I need to be able to crossreference that text with the original as the munged up text has changes i need to make to the original.

Thanks for all the help!

[zwon]

You can start with Text::Levenshtein.

[Fletch]

String::Approx? If that's not what you want the google fodder "Levenshtein distance" is probably a good starting point.

[jethro]

Usually the levenshtein distance is used for this. Search CPAN for the name and you will find Text::Levenshtein and Text::LevenshteinXS. And there are other algorithms, Text::WagnerFischer and Text::PhraseDistance

[planetscape]

I believe there are also many interesting links referenced in Brainstorming session: detecting plagiarism.

[afoken]

A negated hint: Don't use soundex. At least not until you know your input fits into the very narrow scope of the soundex algorithm.

Alexander

[Ratazong]

You may also be interested in the following thread: Fuzzy text matching... again, especially in the node by Limbic~Region.

HTH, Rata