How do I find the percentage of similarity between strings?

[Superman]

I have 2 strings which should be identical:

$PDBline;
$XMLline;

and I test for this normally:

if ($PDBline eq $XMLline){etc...}

However in the event that they dont match I want to do a percentage comparison where I match each word in $PDBline against the words in $XMLline.

So say I want a 75 % similarity between the 2 strings...

my $threshold = 0.75;

my $counter = 0;

foreach my $word ( split ' ', $PDBline ) {
  $counter++ if $XMLline =~ /$word/;
}

Does this seem right or is there an easier way to do this??

Originally posted as a Categorized Question.

[I0]

use String::Approx 'adistr';

[The Mad Hatter]

String::Approx is good as someone else said. So is Text::Levenshtein. With the values you can get from both modules, you should be able to do some math to get a percentage.

You might also want to read Levenshtein distance: calculating similarity of strings.

[chicks]

I'm personally quite fond of Algorithm::Diff. It's line-oriented and you asked for word-oriented, so it might not work for what you're trying to do precisely.

[Superman]

thanks!!!

Originally posted as a Categorized Answer.

[Anonymous Monk]

how do find a percentage?

Originally posted as a Categorized Answer.