• NAME
• SYNOPSIS
• DESCRIPTION
• CONSTRUCTORS
  • new()
• METHODS
  • length_norm(num_tokens)
• INHERITANCE
• COPYRIGHT AND LICENSE

NAME

KinoSearch::Index::Similarity - Judge how well a document matches a query.

SYNOPSIS

    package MySimilarity;

    sub length_norm { return 1.0 }    # disable length normalization

    package MyFullTextType;
    use base qw( KinoSearch::Plan::FullTextType );

    sub make_similarity { MySimilarity->new }

DESCRIPTION

After determining whether a document matches a given query, a score must be calculated which indicates how well the document matches the query. The Similarity class is used to judge how "similar" the query and the document are to each other; the closer the resemblance, they higher the document scores.

The default implementation uses Lucene's modified cosine similarity measure. Subclasses might tweak the existing algorithms, or might be used in conjunction with custom Query subclasses to implement arbitrary scoring schemes.

Most of the methods operate on single fields, but some are used to combine scores from multiple fields.

CONSTRUCTORS

new()

    my $sim = KinoSearch::Index::Similarity->new;

Constructor. Takes no arguments.

METHODS

length_norm(num_tokens)

Dampen the scores of long documents.

After a field is broken up into terms at index-time, each term must be assigned a weight. One of the factors in calculating this weight is the number of tokens that the original field was broken into.

Typically, we assume that the more tokens in a field, the less important any one of them is -- so that, e.g. 5 mentions of "Kafka" in a short article are given more heft than 5 mentions of "Kafka" in an entire book. The default implementation of length_norm expresses this using an inverted square root.

However, the inverted square root has a tendency to reward very short fields highly, which isn't always appropriate for fields you expect to have a lot of tokens on average.

INHERITANCE

KinoSearch::Index::Similarity isa KinoSearch::Object::Obj.

COPYRIGHT AND LICENSE

Copyright 2005-2010 Marvin Humphrey

This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.