view release on metacpan or  search on metacpan

lib/AI/DecisionTree.pm  view on Meta::CPAN

use strict;

package AI::DecisionTree;
{
  $AI::DecisionTree::VERSION = '0.11';
}

use AI::DecisionTree::Instance;
use Carp;
use vars qw(@ISA);


sub new {
  my $package = shift;
  return bless {
		noise_mode => 'fatal',
		prune => 1,
		purge => 1,
		verbose => 0,
		max_depth => 0,
		@_,
		nodes => 0,
		instances => [],
		name_gen => 0,
	       }, $package;
}

sub nodes      { $_[0]->{nodes} }
sub noise_mode { $_[0]->{noise_mode} }
sub depth      { $_[0]->{depth} }

sub add_instance {
  my ($self, %args) = @_;
  croak "Missing 'attributes' parameter" unless $args{attributes};
  croak "Missing 'result' parameter" unless defined $args{result};
  $args{name} = $self->{name_gen}++ unless exists $args{name};
  
  my @attributes;
  while (my ($k, $v) = each %{$args{attributes}}) {
    $attributes[ _hlookup($self->{attributes}, $k) ] = _hlookup($self->{attribute_values}{$k}, $v);
  }
  $_ ||= 0 foreach @attributes;
  
  push @{$self->{instances}}, AI::DecisionTree::Instance->new(\@attributes, _hlookup($self->{results}, $args{result}), $args{name});
}

sub _hlookup {
  $_[0] ||= {}; # Autovivify as a hash
  my ($hash, $key) = @_;
  unless (exists $hash->{$key}) {
    $hash->{$key} = 1 + keys %$hash;
  }
  return $hash->{$key};
}

sub _create_lookup_hashes {
  my $self = shift;
  my $h = $self->{results};
  $self->{results_reverse} = [ undef, sort {$h->{$a} <=> $h->{$b}} keys %$h ];
  
  foreach my $attr (keys %{$self->{attribute_values}}) {
    my $h = $self->{attribute_values}{$attr};
    $self->{attribute_values_reverse}{$attr} = [ undef, sort {$h->{$a} <=> $h->{$b}} keys %$h ];
  }
}

sub train {
  my ($self, %args) = @_;
  if (not @{ $self->{instances} }) {
    croak "Training data has been purged, can't re-train" if $self->{tree};
    croak "Must add training instances before calling train()";
  }
  
  $self->_create_lookup_hashes;
  local $self->{curr_depth} = 0;