Help language development. Donate to The Perl Foundation

Algorithm::NaiveBayes cpan:TITSUKI last updated on 2018-04-01

Algorithm-NaiveBayes-0.0.3/

Build Status

NAME

Algorithm::NaiveBayes - A Perl 6 Naive Bayes classifier implementation

SYNOPSIS

EXAMPLE1

use Algorithm::NaiveBayes;

my $nb = Algorithm::NaiveBayes.new(solver => Algorithm::NaiveBayes::Multinomial);
$nb.add-document("Chinese Beijing Chinese", "China");
$nb.add-document("Chinese Chinese Shanghai", "China");
$nb.add-document("Chinese Macao", "China");
$nb.add-document("Tokyo Japan Chinese", "Japan");
$nb.train();
my @result = $nb.predict("Chinese Chinese Chinese Tokyo Japan");
@result.say; # [China => -8.10769031284391 Japan => -8.90668134500126]

EXAMPLE2

use Algorithm::NaiveBayes;

my $nb = Algorithm::NaiveBayes.new(solver => Algorithm::NaiveBayes::Bernoulli);
$nb.add-document("Chinese Beijing Chinese", "China");
$nb.add-document("Chinese Chinese Shanghai", "China");
$nb.add-document("Chinese Macao", "China");
$nb.add-document("Tokyo Japan Chinese", "Japan");
$nb.train();
my @result = $nb.predict("Chinese Chinese Chinese Tokyo Japan");
@result.say; # [Japan => -3.81908500976888 China => -5.26217831993216]

DESCRIPTION

Algorithm::NaiveBayes is a Perl 6 Naive Bayes classifier implementation.

CONSTRUCTOR

my $nb = Algorithm::NaiveBayes.new(); # default solver is Multinomial
my $nb = Algorithm::NaiveBayes.new(%options);

OPTIONS

METHODS

add-document

multi method add-document(%attributes, Str $label)
multi method add-document(Str @words, Str $label)
multi method add-document(Str $text, Str $label)

Adds a document used for training. %attributes is the key-value pair, where key is the word and value is the frequency of occurrence of the word in the document. @words is the bag-of-words. The bag-of-words is represented as a multiset of words occurrence in the document. $text is the plain text of the document. It will be splitted by whitespaces and processed as the bag-of-words internally.

train

Starts the training.

predict

multi method predict(Str $text)
multi method predict(%attributes)

Returns the log conditional a-posterior probabilities for each class. The resulting list is sorted by descending order of probability. You must call the train method before the prediction.

AUTHOR

titsuki titsuki@cpan.org

COPYRIGHT AND LICENSE

Copyright 2016 titsuki

This library is free software; you can redistribute it and/or modify it under the Artistic License 2.0.

This algorithm is from Manning, Christopher D., Prabhakar Raghavan, and Hinrich Schutze. Introduction to information retrieval. Vol. 1. No. 1. Cambridge: Cambridge university press, 2008.