NAME

Text::Bayon - Handling module for the clustering tool 'Bayon'

SYNOPSIS

  use Text::Bayon;
  use Data::Dumper;

  my $bayon = Text::Bayon->new;
  
  my $input_data = {
        document_id1 => { 
                key1_1 => "value1_1",
                key1_2 => "value1_2",
                key1_3 => "value1_3",
        },
        document_id2 => { 
                key2_1 => "value2_1",
                key2_2 => "value2_2",
                key2_3 => "value2_3",
        },
                .
                .
                .
  };

  my $output = $bayon->clustering($input_data);
  
  print Dumper $output;

  #$output is ... 
  #{
  #  cluster1 => [ document_id, $document_id], 
  #  cluster2 => [ document_id, $document_id], 
  #  cluster2 => [ document_id, $document_id, $document_id], 
  #             .
  #             .
  #             .
  #} 

  #-----------
  # give 'point' option, you can get the data below format.
  #

  my $options = { point => 1 };
  my $output = $bayon->clustering( $input_data, $options );

  print Dumper $output;

  #$output is ... 
  #{
  #  cluster1 => [ { document_id => score}, {$document_id => score} ], 
  #  cluster2 => [ { document_id => score}, {$document_id => score} ], 
  #  cluster2 => [ { document_id => score}, { document_id => score } , 
  #             { document_id => score }, { document_id => score } ], 
  #             .
  #             .
  #             .
  #} 
  
  #-----------
  # set 'clvector' option true, you can get clvector data, too. 
  #

  my $options = { clvector => 1 };
  my ( $output, $clvector ) = $bayon->clustering( $input_data, $options );

  #-----------
  # if you set outfiles as 3rd argument, it restricts returning data and out to files.
  #

  my $options = { clvector => 1 };
  my $outfiles = {
    output   => 'output.tsv',
    clvector => 'centroid.tsv',
  };

  $bayon->clustering( $input_data, $options, $outfiles );

DESCRIPTION

Text::Bayon is handling module for the clustering tool 'Bayon'.

Bayon is a simple and fast hard-clustering tool.

Bayon supports Repeated Bisection clustering and K-means clustering.

I think Bayon is an excellent software for Data-Mining-Peoples!

If you want to know and install Bayon, see the Bayon's maual. ( http://code.google.com/p/bayon/ )

METHODS

new(%conf)

  %conf = (
      bayon_path => '/usr/local/bin/bayon', # optional
      dry_run    => 1, # optional
  );

clustering( $input, $options, $outfiles )

  $input = I< hashref | filename | filehandle >; # required

  $options = {
      number        => I<num>,           # optional
      limit         => I<num>,           # optional, default 1.5 
      point         => 1,                # optional
      clvector      => I< 1 | 0 >,       # optional
      clvector_size => I<num>,           # optional
      method        => I< rb | kmeans >, # optional, default 'rb'
      seed          => num ,             # optional
  };

  $outfiles = {
      output   => I< filename >, # optional
      clvector => I< filename >, # optional
  }

classify($args)

  $input = I< hashref | filename | filehandle >; # required

  $options = {
      classify      => I< filename >, # required
      inv_keys      => I< num >,      # optional, default 20
      inv_size      => I< num >,      # optional, default 100
      classify_size => I< num >,      # optional, default 20
  };

  $outfiles = {
      output   => I< filename >, # optional
  }

AUTHOR

Takeshi Miki <t.miki@nttr.co.jp>

( Bayon's AUTHOR is Mizuki Fujisawa <fujisawa@bayon.cc> )

LICENSE

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

SEE ALSO

http://code.google.com/p/bayon/