Hi there,
I have the following script for computing similarity between two sentences.
#!/usr/bin/perl -w
use strict;
my $candidate = "it is not probable that it is the end";
my $reference = "but it is unlikely that this is the end";
my @candidate_words = split (/\W+/, $candidate);
my $candidate_count=@candidate_words;
my %cand_histogram;
my %ref_histogram;
my $valHolder="";
our $res=0;
$cand_histogram{$_}++ foreach (split(/\s+/,$candidate));
$ref_histogram{$_}++ foreach (split(/\s+/,$reference));
my %seen;
printf "%-6s %-10s %-10s %10s\n", 'Key','Candidate','Reference','MinCo
+unt';
foreach my $key ( sort {
$seen{$b} <=> $seen{$a} #descending word cnt
or
$a cmp $b #alphabetic otherwise
}
grep {!$seen{$_}++} # each key just once,
# but count 'em also!
(keys %cand_histogram)
)
{
#$valHolder=$cand_histogram{$key};
$valHolder=0;
if($cand_histogram{$key} && $ref_histogram{$key}){
$valHolder=$cand_histogram{$key};
if($cand_histogram{$key} >= $ref_histogram{$key}){
$valHolder=$ref_histogram{$key};
}
}
$res+=$valHolder;
printf "%-6s %-10s %-10s %10s\n", $key,
$cand_histogram{$key}||='0',
$ref_histogram{$key} ||='0',$valHolder;
}
my $BS=$res/$candidate_count;
print "The calculated bleu Score is:$BS\n";
Here i have computed the similarity score between two sentences by taking unigrams(i.e individual words from sentences).
Now, i want to turn this same script to do the same computation but with n-gram(with n=3 i.e taking three words together through out the computation).
for eg. in unigram computation if i have a sentence "it is not probably that it is the end". then in n-gram words should be in the form
it is not
is not probably
not probably that
that it is
it is the
is the end
In this way i want to compute the similarity score between the two sentences..
any help please
Thanks..
Output:
Key Candidate Reference MinCount
end 1 1 1
is 2 2 2
it 2 1 1
not 1 0 0
probable 1 0 0
that 1 1 1
the 1 1 1
The calculated bleu Score is:0.666666666666667
-
Are you posting in the right place? Check out Where do I post X? to know for sure.
-
Posts may use any of the Perl Monks Approved HTML tags. Currently these include the following:
<code> <a> <b> <big>
<blockquote> <br /> <dd>
<dl> <dt> <em> <font>
<h1> <h2> <h3> <h4>
<h5> <h6> <hr /> <i>
<li> <nbsp> <ol> <p>
<small> <strike> <strong>
<sub> <sup> <table>
<td> <th> <tr> <tt>
<u> <ul>
-
Snippets of code should be wrapped in
<code> tags not
<pre> tags. In fact, <pre>
tags should generally be avoided. If they must
be used, extreme care should be
taken to ensure that their contents do not
have long lines (<70 chars), in order to prevent
horizontal scrolling (and possible janitor
intervention).
-
Want more info? How to link
or How to display code and escape characters
are good places to start.