=head1 NAME

KinoSearch::Docs::IRTheory - Crash course in information retrieval.

=head1 ABSTRACT

Just enough Information Retrieval theory to find your way around KinoSearch.

=head1 Terminology

KinoSearch uses some terminology from the field of information retrieval which
may be unfamiliar to many users.  "Document" and "term" mean pretty much what
you'd expect them to, but others such as "posting" and "inverted index" need a
formal introduction:

=over

=item *

I<document> - An atomic unit of retrieval.

=item *

I<term> - An attribute which describes a document.

=item *

I<posting> - One term indexing one document.

=item *

I<term list> - The complete list of terms which describe a document.

=item *

I<posting list> - The complete list of documents which a term indexes.

=item *

I<inverted index> - A data structure which maps from terms to documents.

=back

Since KinoSearch is a practical implementation of IR theory, it loads these
abstract, distilled definitions down with useful traits.  For instance,
a "term" need not be associated with a field name; it doesn't even need to be
made out of text.  However, it is convenient to define it that way when trying
to build a text search engine library.

Similarly, a "posting" in its most rarefied form is simply a term-document
pairing; in KinoSearch, the class L<KinoSearch::Posting::MatchPosting> fills
this role.  However, by associating additional information with a posting like
the number of times the term occurs in the document, we can turn it into a
L<ScorePosting|KinoSearch::Posting::ScorePosting>, making it possible to rank
documents by relevance rather than just list documents which happen to match
in no particular order.

=head1 TF/IDF ranking algorithm

KinoSearch uses a variant of the well-established "Term Frequency / Inverse
Document Frequency" weighting scheme.  A thorough treatment of TF/IDF is too
ambitious for our present purposes, but in a nutshell, it means that...

=over

=item

in a search for C<skate park>, documents which score well for the comparatively
rare term C<skate> will rank higher than documents which score well for the
common term C<park>.  

=item

a 10-word text which has one occurrence each of both C<skate> and C<park> will
rank higher than a 1000-word text which also contains one occurrence of each.

=back

A web search for "tf idf" will turn up many excellent explanations of the
algorithm.

=head1 COPYRIGHT

Copyright 2007 Marvin Humphrey

=head1 LICENSE, DISCLAIMER, BUGS, etc.

See L<KinoSearch> version 0.20.

=cut