Random Decision Tree
---------

Programmed by:
        
        Kun Zhang
        Tulane University
	EECS Dept.
	New Orleans, LA, 70118
	USA

Based on J.R. Quinlan C4.5V8.

Algorithm by:
         Wei Fan
         IBM T.J. Watson Center
         Hawthorne, NY 10532

LICENSING TERMS 

This program is granted free of charge for research and education
purposes. However you must obtain a license from the author to use it
for commercial purposes.

Scientific results produced using the software provided shall
acknowledge the use of RDT. Please cite as  

      1. "Learning through Changes: An Empirical Study of Dynamic Behaviors 
       of Probability Estimation Trees" 
       Kun Zhang, Zujia Xu, Jing Peng, and Bill Buckles
       ICDM 2005,

      2. "Wei Fan, Haixun Wang, Philip S. Yu, Sheng Ma: 
       Is random model better? On its accuracy and efficiency. 
       ICDM 2003: 51-58
      
For the details of this algorithm, please refer paper 2.

The software must not be modified and distributed without prior
permission of the author.

By using RDT you agree to the licensing terms.


NO WARRANTY 

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.

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, 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.


Installation
------------

To install RDT, first Create a new directory:

mkdir rdt

Move rdt.tar.gz to this directory and unpack it with

gunzip -c rdt.tar.gz | tar xvf -

Now execute

make or make all

which compiles the system and creates the executable

rdt

then you are ready to go.


How to use
-----------
Several options:

1. to train HALF rdts just with a training set "example.data"

rdt -f example

this will build HALF rdts on this above set. Building Half trees is
the default setting.


2. to train FULL rdts with this training set

rdt -f example -g

this will build Full rdts on the above set

3. to verify HALF rdts performance on test set

rdt -f example -u

4. to build 100 trees(default 10 trees) for example 3 as well as 
random seed changed to 3456 (default 2005)

rdt -f example -u -t 100 -i 3456

5. If test set is available, depends on Full or Half tree is built, 
a probability file named as ".rdh" or ".rdf" 
will be automatically generated after each run. 
It includes the class probability estimation 
for each test instance. 



Format of data set
-------------

format of data set follows the exact requirement of C4.5. 
sample data sets can be found in folder Example_Data.