AI-DecisionTree
view release on metacpan - search on metacpan
view release on metacpan or search on metacpan
Revision history for Perl extension AI::DecisionTree.
0.11 - Fri Mar 2 21:26:51 CST 2012
- Fixed typo in docs - 'leaf_color' should be 'leaf_colors'. [spotted
by Julio Fraire]
- Silence a couple of Perl::Critic warnings
- Convert to Dist::Zilla
0.09 - Sat Jul 26 23:09:44 2008
- Make a new release so that a META.yml file is generated. [Suggested
by Alexandr Ciornii]
- as_graphviz(%args) wasn't passing %args through properly to the
GraphViz->new() method. Fixed. [Ian Knopke]
- get_result() can now take a callback subroutine, rather than just a
hash reference, to check the attributes of the given instance.
This allows lazy instance checking.
0.08 Mon Jul 7 18:01:16 CDT 2003
- Added a 'leaf_color' parameter for making GraphViz objects more
colorful.
0.07 Fri Jun 6 10:37:51 CDT 2003
- Created tests for the set_results() and copy_instances() methods.
- Added documentation for as_graphviz() and increased the information
contained in the GraphViz object.
- Added the ability to limit the absolute depth of the tree when
training.
0.06 Wed Sep 18 13:59:24 EST 2002
- Fixed an XS memory leak that was afflicting all training instances.
Added tests to make sure leak stays plugged.
- Added the 'purge' and 'verbose' parameters to new().
- add_instance() now accepts a 'name' parameter.
- Users can now control whether training instances are purged after
training, using the 'purge' parameter to new() and/or the
do_purge() method.
- Added the set_results() and copy_instances() methods, which let you
re-use training instances from one tree to another.
- Added the instances() and purge() accessor methods.
0.05 Thu Sep 12 01:22:34 AEST 2002
- Fixed a concurrency problem that occurred when making more than one
decision tree. All tree data is now stored as member data, not
class data.
- DecisionTree.pm is now pure-perl again (though Instance.pm still
has an XS component).
- Fixed a one-off bug in the Instance.xs code that could create
garbage data.
- Handles "sparse" data better. Sparse data means that every
attribute doesn't have to be defined for every training/test
instance. This can now be a meaningful property - the absence of a
value is currently equivalent to a special "<undef>" value.
- Don't trigger warnings when undefined attribute values are
encountered (as happens with sparse data).
- Added documentation for the 'prune' parameter to new()
- More consistent with memory allocation in Instance.xs - uses the
perl memory macros/functions from `perldoc perlclib` instead of raw
malloc/realloc/free.
- Catches possible infinite loop situations when growing the tree
(which shouldn't usually happen, but other mistakes can cause it)
- The Instance class now has a numeric-only interface, without string
translation. String translation is done in the main DecisionTree
class. This isn't really a user-visible change.
0.04 Wed Sep 4 19:52:23 AEST 2002
- Now uses regular XS instead of Inline for the C code parts. [patch
by Matt Sergeant]
- Converted the inner loop of the best_attr() method to C code,
because it was spending a lot of time in accessor methods for the C
structures it was using. Don't worry, I'm not going C-crazy. I
won't be making many (any?) more of these kinds of changes, but
these ones were probably necessary.
- Removed a bit of debugging code that I left in for 0.03.
0.03 Mon Sep 2 11:41:18 AEST 2002
- Added a 'prune' parameter to new(), which controls whether the tree
will be pruned after training. This is usually a good idea, so the
default is to prune. Currently we prune using a simple
minimum-description-length criterion.
- Training instances are now represented using a C struct rather than
a Perl hash. This can dramatically reduce memory usage, though it
doesn't have much effect on speed. Note that Inline.pm is now
required.
- The list of instances is now deleted after training, since it's no
longer needed.
- Small speedup to the train() method, achieved by less copying of data.
- If get_result() is called in a list context, it now returns a list
containing the assigned result, a "confidence" score (tentative,
subject to change), and the tree depth of the leaf this instance
ended up at.
- Internally, each node in the tree now contains information about
how many training examples contributed to training this node, and
what the distribution of their classes was.
- Added an as_graphviz() method, which will help visualize trees.
They're not terribly pretty graphviz objects yet, but they're
visual.
0.02 Sat Aug 10, 2002 21:02 AEST
- Added support for noisy data, currently by picking the best (most
common) result when noise is encountered. See the 'noise_mode'
parameter to new().
- Added the rule_tree() method, which returns a data structure
representing the tree. [James Smith]
- Significantly sped up the train() method, especially for large data
sets.
- The get_result() method is no longer implemented recursively, which
simplifies it and speeds it up.
- Reformatted the documentation and added a TO DO section.
- Documented the nodes() method.
0.01 Sat Jun 8 12:45:03 2002
- original version; created by h2xs 1.21 with options
-XA -n AI::DecisionTree
This is the Perl distribution AI-DecisionTree.
Installing AI-DecisionTree is straightforward.
## Installation with cpanm
If you have cpanm, you only need one line:
% cpanm AI::DecisionTree
If you are installing into a system-wide directory, you may need to pass the
"-S" flag to cpanm, which uses sudo to install the module:
% cpanm -S AI::DecisionTree
## Installing with the CPAN shell
Alternatively, if your CPAN shell is set up, you should just be able to do:
% cpan AI::DecisionTree
## Manual installation
As a last resort, you can manually install it. Download the tarball, untar it,
then build it:
% perl Makefile.PL
% make && make test
Then install it:
% make install
If you are installing into a system-wide directory, you may need to run:
% sudo make install
## Documentation
AI-DecisionTree documentation is available as POD.
You can run perldoc from a shell to read the documentation:
% perldoc AI::DecisionTree
Instance/Instance.pm view on Meta::CPAN
1;
__END__
=head1 NAME
AI::DecisionTree::Instance - C-struct wrapper for training instances
=head1 SYNOPSIS
use AI::DecisionTree::Instance;
my $i = new AI::DecisionTree::Instance([3,5], 7, 'this_instance');
$i->value_int(0) == 3;
$i->value_int(1) == 5;
$i->result_int == 7;
=head1 DESCRIPTION
This class is just a simple Perl wrapper around a C struct embodying a
single training instance. Its purpose is to reduce memory usage. In
a "typical" training set with about 1000 instances, memory usage can
be reduced by about a factor of 5 (from 43.7M to 8.2M in my test
program).
A fairly tight loop is also implemented that helps speed up the
Instance/Instance.xs view on Meta::CPAN
extern "C" {
#endif
#include "EXTERN.h"
#include "perl.h"
#include "XSUB.h"
#ifdef __cplusplus
}
#endif
typedef struct {
char *name;
int result;
int num_values;
int *values;
} Instance;
MODULE = AI::DecisionTree::Instance PACKAGE = AI::DecisionTree::Instance
PROTOTYPES: DISABLE
Instance *
new (class, values_ref, result, name)
char * class
SV * values_ref
int result
char * name
CODE:
{
int i;
Instance* instance;
AV* values = (AV*) SvRV(values_ref);
New(0, instance, 1, Instance);
instance->name = savepv(name);
instance->result = result;
instance->num_values = 1 + av_len(values);
New(0, instance->values, instance->num_values, int);
for(i=0; i<instance->num_values; i++) {
instance->values[i] = (int) SvIV( *av_fetch(values, i, 0) );
}
RETVAL = instance;
}
OUTPUT:
RETVAL
char *
name (instance)
Instance* instance
CODE:
{
RETVAL = instance->name;
}
OUTPUT:
RETVAL
void
set_result (instance, result)
Instance* instance
int result
CODE:
{
instance->result = result;
}
void
set_value (instance, attribute, value)
Instance* instance
int attribute
int value
PPCODE:
{
int *new_values;
int i;
if (attribute >= instance->num_values) {
if (!value) return; /* Nothing to do */
printf("Expanding from %d to %d places\n", instance->num_values, attribute);
Renew(instance->values, attribute, int);
if (!instance->values)
croak("Couldn't grab new memory to expand instance");
for (i=instance->num_values; i<attribute-1; i++)
instance->values[i] = 0;
instance->num_values = 1 + attribute;
}
instance->values[attribute] = value;
}
int
value_int (instance, attribute)
Instance * instance
int attribute
CODE:
{
if (attribute >= instance->num_values) {
RETVAL = 0;
}
else {
RETVAL = instance->values[attribute];
}
}
OUTPUT:
RETVAL
int
result_int (instance)
Instance * instance
CODE:
{
RETVAL = instance->result;
}
OUTPUT:
RETVAL
void
DESTROY (instance)
Instance * instance
PPCODE:
{
Safefree(instance->name);
Safefree(instance->values);
Safefree(instance);
}
int
tally (pkg, instances_r, tallies_r, totals_r, attr)
char * pkg
SV * instances_r
SV * tallies_r
SV * totals_r
int attr
CODE:
{
AV *instances = (AV*) SvRV(instances_r);
HV *tallies = (HV*) SvRV(tallies_r);
HV *totals = (HV*) SvRV(totals_r);
I32 top = av_len(instances);
int num_undef = 0;
I32 i, v;
SV **instance_r, **hash_entry, **sub_hash_entry;
Instance *instance;
for (i=0; i<=top; i++) {
instance_r = av_fetch(instances, i, 0);
instance = (Instance *) SvIV(SvRV(*instance_r));
v = attr < instance->num_values ? instance->values[attr] : 0;
/* if (!v) { num_undef++; continue; } */
/* $totals{$v}++ */
hash_entry = hv_fetch(totals, (char *)&v, sizeof(I32), 1);
if (!SvIOK(*hash_entry)) sv_setiv(*hash_entry, 0);
sv_setiv( *hash_entry, 1+SvIV(*hash_entry) );
/* $tallies{$v}{$_->result_int}++ */
hash_entry = hv_fetch(tallies, (char *)&v, sizeof(I32), 0);
if (!hash_entry) {
hash_entry = hv_store(tallies, (char *)&v, sizeof(I32), newRV_noinc((SV*) newHV()), 0);
}
sub_hash_entry = hv_fetch((HV*) SvRV(*hash_entry), (char *)&(instance->result), sizeof(int), 1);
if (!SvIOK(*sub_hash_entry)) sv_setiv(*sub_hash_entry, 0);
sv_setiv( *sub_hash_entry, 1+SvIV(*sub_hash_entry) );
}
RETVAL = num_undef;
/* Old code:
foreach (@$instances) {
my $v = $_->value_int($all_attr->{$attr});
next unless $v;
$totals{ $v }++;
$tallies{ $v }{ $_->result_int }++;
}
*/
}
OUTPUT:
RETVAL
Instance/Makefile.PL view on Meta::CPAN
use ExtUtils::MakeMaker;
my $module = 'AI::DecisionTree::Instance';
my ($name, $dir);
($name = $module) =~ s/.*::(\w+)/$1.pm/;
($dir = $module) =~ s/::/-/g;
WriteMakefile
(
'NAME' => $module,
'VERSION_FROM' => $name, # finds $VERSION
);
Instance/t/02-leaktest.t view on Meta::CPAN
#!/usr/bin/perl
use Test;
BEGIN { plan tests => 4 }
use AI::DecisionTree::Instance;
ok(1);
my $x = 0;
{
local *{"AI::DecisionTree::Instance::DESTROY"} = sub { $x = 1 };
{
my $i = new AI::DecisionTree::Instance([3,4], 4, "foo");
ok $x, 0;
}
ok $x, 1;
}
ok $x, 1;
Instance/typemap view on Meta::CPAN
TYPEMAP
Instance * WRAPPED_C_OBJECT
INPUT
WRAPPED_C_OBJECT
$var = (($type)SvIV(SvRV($arg)))
OUTPUT
WRAPPED_C_OBJECT
/* 'class' must be an argument to any function that returns this datatype */
sv_setref_pv($arg, class, (void*)$var);
This software is copyright (c) 2012 by Ken Williams.
This is free software; you can redistribute it and/or modify it under
the same terms as the Perl 5 programming language system itself.
Terms of the Perl programming language system itself
a) the GNU General Public License as published by the Free
Software Foundation; either version 1, or (at your option) any
later version, or
b) the "Artistic License"
--- The GNU General Public License, Version 1, February 1989 ---
This software is Copyright (c) 2012 by Ken Williams.
This is free software, licensed under:
The GNU General Public License, Version 1, February 1989
GNU GENERAL PUBLIC LICENSE
Version 1, February 1989
Copyright (C) 1989 Free Software Foundation, Inc.
51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
Preamble
The license agreements of most software companies try to keep users
at the mercy of those companies. By contrast, our General Public
License is intended to guarantee your freedom to share and change free
software--to make sure the software is free for all its users. The
General Public License applies to the Free Software Foundation's
software and to any other program whose authors commit to using it.
You can use it for your programs, too.
When we speak of free software, we are referring to freedom, not
price. Specifically, the General Public License is designed to make
sure that you have the freedom to give away or sell copies of free
software, that you receive source code or can get it if you want it,
that you can change the software or use pieces of it in new free
programs; and that you know you can do these things.
To protect your rights, we need to make restrictions that forbid
anyone to deny you these rights or to ask you to surrender the rights.
These restrictions translate to certain responsibilities for you if you
distribute copies of the software, or if you modify it.
For example, if you distribute copies of a such a program, whether
gratis or for a fee, you must give the recipients all the rights that
you have. You must make sure that they, too, receive or can get the
source code. And you must tell them their rights.
We protect your rights with two steps: (1) copyright the software, and
(2) offer you this license which gives you legal permission to copy,
distribute and/or modify the software.
Also, for each author's protection and ours, we want to make certain
that everyone understands that there is no warranty for this free
software. If the software is modified by someone else and passed on, we
want its recipients to know that what they have is not the original, so
that any problems introduced by others will not reflect on the original
authors' reputations.
The precise terms and conditions for copying, distribution and
modification follow.
GNU GENERAL PUBLIC LICENSE
TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
0. This License Agreement applies to any program or other work which
contains a notice placed by the copyright holder saying it may be
distributed under the terms of this General Public License. The
"Program", below, refers to any such program or work, and a "work based
on the Program" means either the Program or any work containing the
Program or a portion of it, either verbatim or with modifications. Each
licensee is addressed as "you".
1. You may copy and distribute verbatim copies of the Program's source
code as you receive it, in any medium, provided that you conspicuously and
appropriately publish on each copy an appropriate copyright notice and
disclaimer of warranty; keep intact all the notices that refer to this
General Public License and to the absence of any warranty; and give any
other recipients of the Program a copy of this General Public License
along with the Program. You may charge a fee for the physical act of
transferring a copy.
2. You may modify your copy or copies of the Program or any portion of
it, and copy and distribute such modifications under the terms of Paragraph
1 above, provided that you also do the following:
a) cause the modified files to carry prominent notices stating that
you changed the files and the date of any change; and
b) cause the whole of any work that you distribute or publish, that
in whole or in part contains the Program or any part thereof, either
with or without modifications, to be licensed at no charge to all
third parties under the terms of this General Public License (except
that you may choose to grant warranty protection to some or all
third parties, at your option).
c) If the modified program normally reads commands interactively when
run, you must cause it, when started running for such interactive use
in the simplest and most usual way, to print or display an
announcement including an appropriate copyright notice and a notice
that there is no warranty (or else, saying that you provide a
warranty) and that users may redistribute the program under these
conditions, and telling the user how to view a copy of this General
Public License.
d) You may charge a fee for the physical act of transferring a
copy, and you may at your option offer warranty protection in
exchange for a fee.
Mere aggregation of another independent work with the Program (or its
derivative) on a volume of a storage or distribution medium does not bring
the other work under the scope of these terms.
3. You may copy and distribute the Program (or a portion or derivative of
it, under Paragraph 2) in object code or executable form under the terms of
Paragraphs 1 and 2 above provided that you also do one of the following:
a) accompany it with the complete corresponding machine-readable
source code, which must be distributed under the terms of
Paragraphs 1 and 2 above; or,
b) accompany it with a written offer, valid for at least three
years, to give any third party free (except for a nominal charge
for the cost of distribution) a complete machine-readable copy of the
corresponding source code, to be distributed under the terms of
Paragraphs 1 and 2 above; or,
c) accompany it with the information you received as to where the
corresponding source code may be obtained. (This alternative is
allowed only for noncommercial distribution and only if you
received the program in object code or executable form alone.)
Source code for a work means the preferred form of the work for making
modifications to it. For an executable file, complete source code means
all the source code for all modules it contains; but, as a special
exception, it need not include source code for modules which are standard
libraries that accompany the operating system on which the executable
file runs, or for standard header files or definitions files that
accompany that operating system.
4. You may not copy, modify, sublicense, distribute or transfer the
Program except as expressly provided under this General Public License.
Any attempt otherwise to copy, modify, sublicense, distribute or transfer
the Program is void, and will automatically terminate your rights to use
the Program under this License. However, parties who have received
copies, or rights to use copies, from you under this General Public
License will not have their licenses terminated so long as such parties
remain in full compliance.
5. By copying, distributing or modifying the Program (or any work based
on the Program) you indicate your acceptance of this license to do so,
and all its terms and conditions.
6. Each time you redistribute the Program (or any work based on the
Program), the recipient automatically receives a license from the original
licensor to copy, distribute or modify the Program subject to these
terms and conditions. You may not impose any further restrictions on the
recipients' exercise of the rights granted herein.
7. The Free Software Foundation may publish revised and/or new versions
of the General Public License from time to time. Such new versions will
be similar in spirit to the present version, but may differ in detail to
address new problems or concerns.
Each version is given a distinguishing version number. If the Program
specifies a version number of the license which applies to it and "any
later version", you have the option of following the terms and conditions
either of that version or of any later version published by the Free
Software Foundation. If the Program does not specify a version number of
the license, you may choose any version ever published by the Free Software
Foundation.
8. If you wish to incorporate parts of the Program into other free
programs whose distribution conditions are different, write to the author
to ask for permission. For software which is copyrighted by the Free
Software Foundation, write to the Free Software Foundation; we sometimes
make exceptions for this. Our decision will be guided by the two goals
of preserving the free status of all derivatives of our free software and
of promoting the sharing and reuse of software generally.
NO WARRANTY
9. BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY
FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN
OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES
PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED
OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS
TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE
PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING,
REPAIR OR CORRECTION.
10. IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING
WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR
REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES,
INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING
OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED
TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY
YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER
PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES.
END OF TERMS AND CONDITIONS
Appendix: How to Apply These Terms to Your New Programs
If you develop a new program, and you want it to be of the greatest
possible use to humanity, the best way to achieve this is to make it
free software which everyone can redistribute and change under these
terms.
To do so, attach the following notices to the program. It is safest to
attach them to the start of each source file to most effectively convey
the exclusion of warranty; and each file should have at least the
"copyright" line and a pointer to where the full notice is found.
<one line to give the program's name and a brief idea of what it does.>
Copyright (C) 19yy <name of author>
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 1, or (at your option)
any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301 USA
Also add information on how to contact you by electronic and paper mail.
If the program is interactive, make it output a short notice like this
when it starts in an interactive mode:
Gnomovision version 69, Copyright (C) 19xx name of author
Gnomovision comes with ABSOLUTELY NO WARRANTY; for details type `show w'.
This is free software, and you are welcome to redistribute it
under certain conditions; type `show c' for details.
The hypothetical commands `show w' and `show c' should show the
appropriate parts of the General Public License. Of course, the
commands you use may be called something other than `show w' and `show
c'; they could even be mouse-clicks or menu items--whatever suits your
program.
You should also get your employer (if you work as a programmer) or your
school, if any, to sign a "copyright disclaimer" for the program, if
necessary. Here a sample; alter the names:
Yoyodyne, Inc., hereby disclaims all copyright interest in the
program `Gnomovision' (a program to direct compilers to make passes
at assemblers) written by James Hacker.
<signature of Ty Coon>, 1 April 1989
Ty Coon, President of Vice
That's all there is to it!
--- The Artistic License 1.0 ---
This software is Copyright (c) 2012 by Ken Williams.
This is free software, licensed under:
The Artistic License 1.0
The Artistic License
Preamble
The intent of this document is to state the conditions under which a Package
may be copied, such that the Copyright Holder maintains some semblance of
artistic control over the development of the package, while giving the users of
the package the right to use and distribute the Package in a more-or-less
customary fashion, plus the right to make reasonable modifications.
Definitions:
- "Package" refers to the collection of files distributed by the Copyright
Holder, and derivatives of that collection of files created through
textual modification.
- "Standard Version" refers to such a Package if it has not been modified,
or has been modified in accordance with the wishes of the Copyright
Holder.
- "Copyright Holder" is whoever is named in the copyright or copyrights for
the package.
- "You" is you, if you're thinking about copying or distributing this Package.
- "Reasonable copying fee" is whatever you can justify on the basis of media
cost, duplication charges, time of people involved, and so on. (You will
not be required to justify it to the Copyright Holder, but only to the
computing community at large as a market that must bear the fee.)
- "Freely Available" means that no fee is charged for the item itself, though
there may be fees involved in handling the item. It also means that
recipients of the item may redistribute it under the same conditions they
received it.
1. You may make and give away verbatim copies of the source form of the
Standard Version of this Package without restriction, provided that you
duplicate all of the original copyright notices and associated disclaimers.
2. You may apply bug fixes, portability fixes and other modifications derived
from the Public Domain or from the Copyright Holder. A Package modified in such
a way shall still be considered the Standard Version.
3. You may otherwise modify your copy of this Package in any way, provided that
you insert a prominent notice in each changed file stating how and when you
changed that file, and provided that you do at least ONE of the following:
a) place your modifications in the Public Domain or otherwise make them
Freely Available, such as by posting said modifications to Usenet or an
equivalent medium, or placing the modifications on a major archive site
such as ftp.uu.net, or by allowing the Copyright Holder to include your
modifications in the Standard Version of the Package.
b) use the modified Package only within your corporation or organization.
c) rename any non-standard executables so the names do not conflict with
standard executables, which must also be provided, and provide a separate
manual page for each non-standard executable that clearly documents how it
differs from the Standard Version.
d) make other distribution arrangements with the Copyright Holder.
4. You may distribute the programs of this Package in object code or executable
form, provided that you do at least ONE of the following:
a) distribute a Standard Version of the executables and library files,
together with instructions (in the manual page or equivalent) on where to
get the Standard Version.
b) accompany the distribution with the machine-readable source of the Package
with your modifications.
c) accompany any non-standard executables with their corresponding Standard
Version executables, giving the non-standard executables non-standard
names, and clearly documenting the differences in manual pages (or
equivalent), together with instructions on where to get the Standard
Version.
d) make other distribution arrangements with the Copyright Holder.
5. You may charge a reasonable copying fee for any distribution of this
Package. You may charge any fee you choose for support of this Package. You
may not charge a fee for this Package itself. However, you may distribute this
Package in aggregate with other (possibly commercial) programs as part of a
larger (possibly commercial) software distribution provided that you do not
advertise this Package as a product of your own.
6. The scripts and library files supplied as input to or produced as output
from the programs of this Package do not automatically fall under the copyright
---
abstract: 'Automatically Learns Decision Trees'
author:
- 'Ken Williams <kwilliams@cpan.org>'
build_requires:
English: 0
Test: 0
Test::More: 0
warnings: 0
configure_requires:
ExtUtils::MakeMaker: 6.30
dynamic_config: 0
generated_by: 'Dist::Zilla version 4.300009, CPAN::Meta::Converter version 2.120351'
license: perl
meta-spec:
url: http://module-build.sourceforge.net/META-spec-v1.4.html
version: 1.4
name: AI-DecisionTree
requires:
Carp: 0
GraphViz: 0
strict: 0
vars: 0
resources:
bugtracker: http://rt.cpan.org/Public/Dist/Display.html?Name=AI-DecisionTree
repository: git://github.com/kenahoo/perl-ai-decisiontree.git
version: 0.11
Makefile.PL view on Meta::CPAN
use strict;
use warnings;
use ExtUtils::MakeMaker 6.30;
my %WriteMakefileArgs = (
"ABSTRACT" => "Automatically Learns Decision Trees",
"AUTHOR" => "Ken Williams <kwilliams\@cpan.org>",
"BUILD_REQUIRES" => {
"English" => 0,
"Test" => 0,
"Test::More" => 0,
"warnings" => 0
},
"CONFIGURE_REQUIRES" => {
"ExtUtils::MakeMaker" => "6.30"
},
"DISTNAME" => "AI-DecisionTree",
"EXE_FILES" => [],
"LICENSE" => "perl",
"NAME" => "AI::DecisionTree",
"PREREQ_PM" => {
"Carp" => 0,
"GraphViz" => 0,
"strict" => 0,
"vars" => 0
},
"VERSION" => "0.11",
"test" => {
"TESTS" => "t/*.t"
}
);
unless ( eval { ExtUtils::MakeMaker->VERSION(6.56) } ) {
my $br = delete $WriteMakefileArgs{BUILD_REQUIRES};
my $pp = $WriteMakefileArgs{PREREQ_PM};
for my $mod ( keys %$br ) {
if ( exists $pp->{$mod} ) {
$pp->{$mod} = $br->{$mod} if $br->{$mod} > $pp->{$mod};
}
else {
$pp->{$mod} = $br->{$mod};
}
}
}
delete $WriteMakefileArgs{CONFIGURE_REQUIRES}
unless eval { ExtUtils::MakeMaker->VERSION(6.52) };
WriteMakefile(%WriteMakefileArgs);
This archive contains the distribution AI-DecisionTree,
version 0.11:
Automatically Learns Decision Trees
This software is copyright (c) 2012 by Ken Williams.
This is free software; you can redistribute it and/or modify it under
the same terms as the Perl 5 programming language system itself.
This file contains message digests of all files listed in MANIFEST,
signed via the Module::Signature module, version 0.68.
To verify the content in this distribution, first make sure you have
Module::Signature installed, then type:
% cpansign -v
It will check each file's integrity, as well as the signature's
validity. If "==> Signature verified OK! <==" is not displayed,
the distribution may already have been compromised, and you should
not run its Makefile.PL or Build.PL.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
SHA1 e01a2ac643b28c7aa7d0d07a8185899a6aa32b13 Changes
eg/example.pl view on Meta::CPAN
#!/usr/bin/perl
use AI::DecisionTree;
my @attributes = qw(outlook temperature humidity wind play_tennis);
my @cases = qw(
sunny hot high weak no
sunny hot high strong no
overcast hot high weak yes
rain mild high weak yes
rain cool normal weak yes
rain cool normal strong no
overcast cool normal strong yes
sunny mild high weak no
sunny cool normal weak yes
rain mild normal weak yes
sunny mild normal strong yes
overcast mild high strong yes
overcast hot normal weak yes
rain mild high strong no
);
my $outcome = pop @attributes;
my $dtree = new AI::DecisionTree;
while (@cases) {
my @values = splice @cases, 0, 1 + scalar(@attributes);
my $result = pop @values;
my %pairs;
@pairs{@attributes} = @values;
$dtree->add_instance(attributes => \%pairs,
result => $result,
);
}
$dtree->train;
my $result;
# Try one of the training examples
$result = $dtree->get_result( attributes => {
outlook => 'rain',
temperature => 'mild',
humidity => 'high',
wind => 'strong',
} );
print "Result 1: $result\n"; # no
# Try a new unseen example
$result = $dtree->get_result( attributes => {
outlook => 'sunny',
temperature => 'hot',
humidity => 'normal',
wind => 'strong',
} );
print "Result 2: $result\n"; # yes
# Show the created tree structure as rules
print map "$_\n", $dtree->rule_statements;
# Will barf on inconsistent data
my $t2 = new AI::DecisionTree;
$t2->add_instance( attributes => { foo => 'bar' },
result => 1 );
$t2->add_instance( attributes => { foo => 'bar' },
result => 0 );
eval {$t2->train};
print "$@\n";
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;
local $self->{max_depth} = $args{max_depth} if exists $args{max_depth};
$self->{depth} = 0;
$self->{tree} = $self->_expand_node( instances => $self->{instances} );
$self->{total_instances} = @{$self->{instances}};
$self->prune_tree if $self->{prune};
$self->do_purge if $self->purge;
return 1;
}
sub do_purge {
my $self = shift;
delete @{$self}{qw(instances attribute_values attribute_values_reverse results results_reverse)};
}
sub copy_instances {
my ($self, %opt) = @_;
croak "Missing 'from' parameter to copy_instances()" unless exists $opt{from};
my $other = $opt{from};
croak "'from' parameter is not a decision tree" unless UNIVERSAL::isa($other, __PACKAGE__);
foreach (qw(instances attributes attribute_values results)) {
$self->{$_} = $other->{$_};
}
$self->_create_lookup_hashes;
}
sub set_results {
my ($self, $hashref) = @_;
foreach my $instance (@{$self->{instances}}) {
my $name = $instance->name;
croak "No result given for instance '$name'" unless exists $hashref->{$name};
$instance->set_result( $self->{results}{ $hashref->{$name} } );
}
}
sub instances { $_[0]->{instances} }
sub purge {
my $self = shift;
$self->{purge} = shift if @_;
return $self->{purge};
}
# Each node contains:
# { split_on => $attr_name,
# children => { $attr_value1 => $node1,
# $attr_value2 => $node2, ... }
# }
# or
# { result => $result }
sub _expand_node {
my ($self, %args) = @_;
my $instances = $args{instances};
print STDERR '.' if $self->{verbose};
$self->{depth} = $self->{curr_depth} if $self->{curr_depth} > $self->{depth};
local $self->{curr_depth} = $self->{curr_depth} + 1;
$self->{nodes}++;
my %results;
$results{$self->_result($_)}++ foreach @$instances;
my @results = map {$_,$results{$_}} sort {$results{$b} <=> $results{$a}} keys %results;
my %node = ( distribution => \@results, instances => scalar @$instances );
foreach (keys %results) {
$self->{prior_freqs}{$_} += $results{$_};
}
if (keys(%results) == 1) {
# All these instances have the same result - make this node a leaf
$node{result} = $self->_result($instances->[0]);
return \%node;
}
# Multiple values are present - find the best predictor attribute and split on it
my $best_attr = $self->best_attr($instances);
croak "Inconsistent data, can't build tree with noise_mode='fatal'"
if $self->{noise_mode} eq 'fatal' and !defined $best_attr;
if ( !defined($best_attr)
or $self->{max_depth} && $self->{curr_depth} > $self->{max_depth} ) {
# Pick the most frequent result for this leaf
$node{result} = (sort {$results{$b} <=> $results{$a}} keys %results)[0];
return \%node;
}
$node{split_on} = $best_attr;
my %split;
foreach my $i (@$instances) {
my $v = $self->_value($i, $best_attr);
push @{$split{ defined($v) ? $v : '<undef>' }}, $i;
}
die ("Something's wrong: attribute '$best_attr' didn't split ",
scalar @$instances, " instances into multiple buckets (@{[ keys %split ]})")
unless keys %split > 1;
foreach my $value (keys %split) {
$node{children}{$value} = $self->_expand_node( instances => $split{$value} );
}
return \%node;
}
sub best_attr {
my ($self, $instances) = @_;
# 0 is a perfect score, entropy(#instances) is the worst possible score
my ($best_score, $best_attr) = (@$instances * $self->entropy( map $_->result_int, @$instances ), undef);
my $all_attr = $self->{attributes};
foreach my $attr (keys %$all_attr) {
# %tallies is correlation between each attr value and result
# %total is number of instances with each attr value
my (%totals, %tallies);
my $num_undef = AI::DecisionTree::Instance::->tally($instances, \%tallies, \%totals, $all_attr->{$attr});
next unless keys %totals; # Make sure at least one instance defines this attribute
my $score = 0;
while (my ($opt, $vals) = each %tallies) {
$score += $totals{$opt} * $self->entropy2( $vals, $totals{$opt} );
}
($best_attr, $best_score) = ($attr, $score) if $score < $best_score;
}
return $best_attr;
}
sub entropy2 {
shift;
my ($counts, $total) = @_;
# Entropy is defined with log base 2 - we just divide by log(2) at the end to adjust.
my $sum = 0;
$sum += $_ * log($_) foreach values %$counts;
return +(log($total) - $sum/$total)/log(2);
}
sub entropy {
shift;
my %count;
$count{$_}++ foreach @_;
# Entropy is defined with log base 2 - we just divide by log(2) at the end to adjust.
my $sum = 0;
$sum += $_ * log($_) foreach values %count;
return +(log(@_) - $sum/@_)/log(2);
}
sub prune_tree {
my $self = shift;
# We use a minimum-description-length approach. We calculate the
# score of each node:
# n = number of nodes below
# r = number of results (categories) in the entire tree
# i = number of instances in the entire tree
# e = number of errors below this node
# Hypothesis description length (MML):
# describe tree: number of nodes + number of edges
# describe exceptions: num_exceptions * log2(total_num_instances) * log2(total_num_results)
my $r = keys %{ $self->{results} };
my $i = $self->{tree}{instances};
my $exception_cost = log($r) * log($i) / log(2)**2;
# Pruning can turn a branch into a leaf
my $maybe_prune = sub {
my ($self, $node) = @_;
return unless $node->{children}; # Can't prune leaves
my $nodes_below = $self->nodes_below($node);
my $tree_cost = 2 * $nodes_below - 1; # $edges_below == $nodes_below - 1
my $exceptions = $self->exceptions( $node );
my $simple_rule_exceptions = $node->{instances} - $node->{distribution}[1];
my $score = -$nodes_below - ($exceptions - $simple_rule_exceptions) * $exception_cost;
#warn "Score = $score = -$nodes_below - ($exceptions - $simple_rule_exceptions) * $exception_cost\n";
if ($score < 0) {
delete @{$node}{'children', 'split_on', 'exceptions', 'nodes_below'};
$node->{result} = $node->{distribution}[0];
# XXX I'm not cleaning up 'exceptions' or 'nodes_below' keys up the tree
}
};
$self->_traverse($maybe_prune);
}
sub exceptions {
my ($self, $node) = @_;
return $node->{exceptions} if exists $node->{exeptions};
my $count = 0;
if ( exists $node->{result} ) {
$count = $node->{instances} - $node->{distribution}[1];
} else {
foreach my $child ( values %{$node->{children}} ) {
$count += $self->exceptions($child);
}
}
return $node->{exceptions} = $count;
}
sub nodes_below {
my ($self, $node) = @_;
return $node->{nodes_below} if exists $node->{nodes_below};
my $count = 0;
$self->_traverse( sub {$count++}, $node );
return $node->{nodes_below} = $count - 1;
}
# This is *not* for external use, I may change it.
sub _traverse {
my ($self, $callback, $node, $parent, $node_name) = @_;
$node ||= $self->{tree};
ref($callback) ? $callback->($self, $node, $parent, $node_name) : $self->$callback($node, $parent, $node_name);
return unless $node->{children};
foreach my $child ( keys %{$node->{children}} ) {
$self->_traverse($callback, $node->{children}{$child}, $node, $child);
}
}
sub get_result {
my ($self, %args) = @_;
croak "Missing 'attributes' or 'callback' parameter" unless $args{attributes} or $args{callback};
$self->train unless $self->{tree};
my $tree = $self->{tree};
while (1) {
if (exists $tree->{result}) {
my $r = $tree->{result};
return $r unless wantarray;
my %dist = @{$tree->{distribution}};
my $confidence = $tree->{distribution}[1] / $tree->{instances};
# my $confidence = P(H|D) = [P(D|H)P(H)]/[P(D|H)P(H)+P(D|H')P(H')]
# = [P(D|H)P(H)]/P(D);
# my $confidence =
# $confidence *= $self->{prior_freqs}{$r} / $self->{total_instances};
return ($r, $confidence, \%dist);
}
my $instance_val = (exists $args{callback} ? $args{callback}->($tree->{split_on}) :
exists $args{attributes}{$tree->{split_on}} ? $args{attributes}{$tree->{split_on}} :
'<undef>');
## no critic (ProhibitExplicitReturnUndef)
$tree = $tree->{children}{ $instance_val }
or return undef;
}
}
sub as_graphviz {
my ($self, %args) = @_;
my $colors = delete $args{leaf_colors} || {};
require GraphViz;
my $g = GraphViz->new(%args);
my $id = 1;
my $add_edge = sub {
my ($self, $node, $parent, $node_name) = @_;
# We use stringified reference names for node names, as a convenient hack.
if ($node->{split_on}) {
$g->add_node( "$node",
label => $node->{split_on},
shape => 'ellipse',
);
} else {
my $i = 0;
my $distr = join ',', grep {$i++ % 2} @{$node->{distribution}};
my %fill = (exists $colors->{$node->{result}} ?
(fillcolor => $colors->{$node->{result}},
style => 'filled') :
()
);
$g->add_node( "$node",
label => "$node->{result} ($distr)",
shape => 'box',
%fill,
);
}
$g->add_edge( "$parent" => "$node",
label => $node_name,
) if $parent;
};
$self->_traverse( $add_edge );
return $g;
}
sub rule_tree {
my $self = shift;
my ($tree) = @_ ? @_ : $self->{tree};
# build tree:
# [ question, { results => [ question, { ... } ] } ]
return $tree->{result} if exists $tree->{result};
return [
$tree->{split_on}, {
map { $_ => $self->rule_tree($tree->{children}{$_}) } keys %{$tree->{children}},
}
];
}
sub rule_statements {
my $self = shift;
my ($stmt, $tree) = @_ ? @_ : ('', $self->{tree});
return("$stmt -> '$tree->{result}'") if exists $tree->{result};
my @out;
my $prefix = $stmt ? "$stmt and" : "if";
foreach my $val (keys %{$tree->{children}}) {
push @out, $self->rule_statements("$prefix $tree->{split_on}='$val'", $tree->{children}{$val});
}
return @out;
}
### Some instance accessor stuff:
sub _result {
my ($self, $instance) = @_;
my $int = $instance->result_int;
return $self->{results_reverse}[$int];
}
sub _delete_value {
my ($self, $instance, $attr) = @_;
my $val = $self->_value($instance, $attr);
return unless defined $val;
$instance->set_value($self->{attributes}{$attr}, 0);
return $val;
}
sub _value {
my ($self, $instance, $attr) = @_;
return unless exists $self->{attributes}{$attr};
my $val_int = $instance->value_int($self->{attributes}{$attr});
return $self->{attribute_values_reverse}{$attr}[$val_int];
}
1;
__END__
=head1 NAME
AI::DecisionTree - Automatically Learns Decision Trees
=head1 VERSION
version 0.11
=head1 SYNOPSIS
use AI::DecisionTree;
my $dtree = new AI::DecisionTree;
# A set of training data for deciding whether to play tennis
$dtree->add_instance
(attributes => {outlook => 'sunny',
temperature => 'hot',
humidity => 'high'},
result => 'no');
$dtree->add_instance
(attributes => {outlook => 'overcast',
temperature => 'hot',
humidity => 'normal'},
result => 'yes');
... repeat for several more instances, then:
$dtree->train;
# Find results for unseen instances
my $result = $dtree->get_result
(attributes => {outlook => 'sunny',
temperature => 'hot',
humidity => 'normal'});
=head1 DESCRIPTION
The C<AI::DecisionTree> module automatically creates so-called
"decision trees" to explain a set of training data. A decision tree
is a kind of categorizer that use a flowchart-like process for
categorizing new instances. For instance, a learned decision tree
might look like the following, which classifies for the concept "play
tennis":
OUTLOOK
/ | \
/ | \
/ | \
sunny/ overcast \rainy
/ | \
HUMIDITY | WIND
/ \ *no* / \
/ \ / \
high/ \normal / \
/ \ strong/ \weak
*no* *yes* / \
*no* *yes*
(This example, and the inspiration for the C<AI::DecisionTree> module,
come directly from Tom Mitchell's excellent book "Machine Learning",
available from McGraw Hill.)
A decision tree like this one can be learned from training data, and
then applied to previously unseen data to obtain results that are
consistent with the training data.
The usual goal of a decision tree is to somehow encapsulate the
t/02-noisy.t view on Meta::CPAN
my $dtree = AI::DecisionTree->new(noise_mode => 'pick_best');
ok $dtree;
my @names = split /, /, <DATA>;
chomp $names[-1];
# Train on first 600 instances
printf "Loading 600 training instances with %d attribute types each\n", scalar @names;
while (<DATA>) {
last unless 2..601;
chomp;
my @values = split /, /, $_;
my $result = pop @values;
my %pairs = map {$names[$_], $values[$_]} 0..$#names;
$dtree->add_instance(attributes => \%pairs,
result => $result,
);
}
print "Building decision tree\n";
$dtree->train;
ok(1);
# Test on rest of data, get at least 80%
print "Testing on remainder of data\n";
my ($good, $bad) = (0,0);
while (<DATA>) {
chomp;
my @values = split /, /, $_;
my $result = pop @values;
my %pairs = map {$names[$_], $values[$_]} 0..$#names;
my ($guess, $confidence) = $dtree->get_result(attributes => \%pairs);
$guess ||= ''; $confidence ||= '';
($guess eq $result ? $good : $bad)++;
#print "$guess : $result : $confidence\n";
}
my $accuracy = $good/($good + $bad);
ok $accuracy > .8;
print "Accuracy=$accuracy\n";
#use YAML; print Dump($dtree->rule_tree);
#print map "$_\n", $dtree->rule_statements;
if (eval "use GraphViz; 1") {
my $graphviz = $dtree->as_graphviz;
ok $graphviz;
if (0) {
# Only works on Mac OS X
my $file = '/tmp/tree2.png';
open my($fh), "> $file" or die "$file: $!";
print $fh $graphviz->as_png;
close $fh;
system('open', $file);
}
} else {
skip("Skipping: GraphViz is not installed", 0);
}
# The following data comes from the "C4.5" software package, in the
# "soybean.data" data file. It is somewhat noisy. I chose it because
# it was a pretty big data set, and because there are published
# results on it that I can compare to. Since the data seemed to be in
# order from most-information to least-information or something in the
# C4.5 distribution, I randomized the order of the instances. Note
# also that I'm treating the '?' value as a normal string value.
t/author-critic.t view on Meta::CPAN
#!perl
BEGIN {
unless ($ENV{AUTHOR_TESTING}) {
require Test::More;
Test::More::plan(skip_all => 'these tests are for testing by the author');
}
}
use strict;
use warnings;
use Test::More;
use English qw(-no_match_vars);
eval "use Test::Perl::Critic";
view release on metacpan - search on metacpan
( run in 1.358 second using v1.00-cache-2.02-grep-82fe00e-cpan-2c419f77a38b )