NAME

AI::subclust - A module to implement substractive clustering algorithm.

SYNOPSYS

    use AI::subclust;

    my @Data  = [[qw/0.3 0.5/],[qw/0.3 0.6/],[qw/0.6 0.8/],[qw/0.02 0.6/]];
    my @bound = [[qw/0 0/],[qw/1 1/]];
    
    my $subC = new AI::subclust(-data=>@Data,-bounds=>@bound);

    my ($CLU,$S) = $subC->calculate();

DESCRIPTION

This module implements a substractive clustering algorithm.

PUBLIC METHODS

The module has the following public methods:

new()

This is the constructor. It have to be defined only -data paramater obligatorily any other has an default value.

$subC = new subclust( -data => @Data, -bounds => @bounds, -ra => 0.5, -rb => 0.6, -acceptLimit => 0.5, -vervose => 0, );

-data

Input data array.

-bounds

Scale array. It is a 2xN dimension array. Use default valueas @max and @min arrays calculated from @Data

-ra

Radio of the hypercube selected for each cluster. Use default value as 0.5

-rb

Radio of rejection for each cluster. Use default value as 0.6

-acceptLimit

Minimal value of Potencial required to be center of cluster. Use default value as 0.5

-vervose

Vevose mode. 1 activated or 0 desactivated. Use default value as 0

calculate()

This method is used to calculate substractive algorithm. Just call method in this way

($CLU,$S) = $subC->calculate();

Method retrieve two array references, following with above example: $CLU represent an cluster position array reference $S represent sigma value array reference of each variable (Matlab style).

AUTHOR

Copyright 2004, Jorge Courett. All rights reserved. This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.

Please send me any comments to: courett@softhome.net