NAME

AI::NeuralNet::Kohonen::Demo::RGB - Colour-based demo

SYNOPSIS

use AI::NeuralNet::Kohonen::Demo::RGB;
$_ = AI::NeuralNet::Kohonen::Demo::RGB->new(
	display_scale => 20,
	display	=> 'hex',
	map_dim	=> 39,
	epochs  => 9,
	table   => "R G B"
              ."1 0 0"
              ."0 1 0"
              ."0 0 1",
);
$_->train;
exit;

DESCRIPTION

A sub-class of AI::NeuralNet::Kohonen that impliments extra methods to make use of TK in a very slow demonstration of how a SOM can collapse a three dimensional space (RGB colour values) into a two dimensional space (the display). See SYNOPSIS.

The only things added are two new fields to supply to the constructor - set display to hex for display as a unified distance matrix, rather than plain grid; set display_scale for the size of the display.

METHOD train

Over-rides the base class to provide TK displays of the map

SEE ALSO

See AI::NeuralNet::Kohonen; AI::NeuralNet::Kohonen::Node;

AUTHOR AND COYRIGHT

This implimentation Copyright (C) Lee Goddard, 2003. All Rights Reserved.

Available under the same terms as Perl itself.