Mark Leighton Fisher >
Set-Jaccard-SimilarityCoefficient-0.5.1 >
Set::Jaccard::SimilarityCoefficient

Module Version: 0.5.1
Calculate the Jaccard Similarity Coefficient.

Set::Jaccard::SimilarityCoefficient - Calculate the Jaccard Similarity Coefficient of 2 sets

# VERSION

$res = Set::Jaccard::SimilarityCoefficient::calc(\@set_a, \@set_b);

OR

my $a = Set::Scalar->new(@set_a); my $b = Set::Scalar->new(@set_b); $res = Set::Jaccard::SimilarityCoefficient::calc($a, $b);

Set::Jaccard::SimilarityCoefficient lets you calculate the Jaccard Similarity Coefficient for either arrayrefs or Set::Scalar objects.

Briefly, the Jaccard Similarity Coefficient is a simple measure of how similar 2 sets are. The calculation is (in pseudo-code):

count(difference(SET-A, SET-B)) / count(union(SET-A, SET-B))

There is a Jaccard Similarity Coefficient routine already in CPAN, but it is specialized for use by Text::NSP. I wanted a generic routine that could be used by anyone so Set::Jaccard::SimilarityCoefficient was born.

calc(A, B) calculates the Jaccard Similarity Coefficient for the arguments A and B. A and B can be either array references or Set::Scalar objects.

new() will complain if A or B is empty, not either a reference to an array, or not a Set::Scalar object.

calc() could theoretically throw DivideByZeroException when the union of the two sets has 0 members. However, that would require set A or set B to have 0 members, which was previously prohibited by the prohibition on empty sets.

This module should work wherever Perl works.

Set::Scalar

None that I know of.

There are no bugs that I know of. Given that this is non-trivial code, there will be bugs.

The types of arguments are limited to either array references or Set::Scalar objects.

Mark Leighton Fisher, <markleightonfisher@gmail.com>

Set::JaccardSimilarityCoefficient is licensed under the same terms as Perl itself.

syntax highlighting: