****************************************************************
* Copyright 1993,1994 : Johns Hopkins University		*
*                       Department of Computer Science		*
****************************************************************
* Contacts : [email protected] (Sreerama K. Murthy)             *
*            [email protected] (Steven Salzberg)              *
*            [email protected] (Simon Kasif)                     *
****************************************************************

***** Notice for readers/users who retrieved this from the JAIR
      online appendix *****

  This source code is supplied "as is" without warranty of any kind, and
  its authors and the Journal of Artificial Intelligence Research (JAIR)
  and JAIR's publishers and distributors, disclaim any and all
  warranties, including but not limited to any implied warranties of
  merchantability and fitness for a particular purpose, and any
  warranties or non infringement.  The user assumes all liability and
  responsibility for use of this source code, and neither the author nor
  JAIR, nor JAIR's publishers and distributors, will be liable for
  damages of any kind resulting from its use.  Without limiting the
  generality of the foregoing, neither the authors, nor JAIR, nor JAIR's
  publishers and distributors, warrant that the Source Code will be
  error-free, will operate without interruption, or will meet the needs
  of the user.

****************************************************************

Welcome to OC1 Version 3.0!

OC1 is a system to construct oblique decision trees from examples.
Oblique decision trees are trees in which each node may contain a
(linear) multivariate test on the attributes of the data.  OC1 also
constructs standard axis-parallel trees, which contain tests of just
one attribute at each node.  Oblique decision trees are a natural
extension to the well-known axis-parallel trees.

As no one decision tree building method (or, for that matter, machine
learning method) is the best for all datasets, we feel that a machine
learning researcher/practitioner should experiment with as many
methods as possible when attempting to solve a problem.  To aid in
this goal, we are making OC1 available in the public domain.  Please
use it, experiment with it, and let us know your questions, comments
or suggestions.  OC1 is intended for non-commercial use only, and you
should feel free to use, copy, and modify it for such purposes.  Any
commercial use of OC1 is strictly prohibited without the express
written consent of the authors.

The OC1 directory has four main components: 
  "gendata" generates artificial datasets, given (optionally) a 
            decision tree; 
  "mktree"  builds decision trees out of data, estimates classification 
            accuracies; 
  "display" displays 2D datasets and/or decision trees as PostScript(R) files. 
  "jair94-paper.ps" contains the PostScript(R) version of our paper:
     S.K. Murthy, S. Kasif, S. Salzberg.
     "A System for Induction of Oblique Decision Trees."
     Journal of Artificial Intelligence Research 2 (1994) 1-33.

If you use the OC1 software in the context of any of your
publications, please reference the above paper.

To install OC1, after "FTP"ing, uncompressing and unarchiving ("tar"ing) 
the software, run the following commands:
      $ make mktree
      $ make gendata
      $ make display 

These commands will create the executable files for the three main
commands available in the OC1 system.  You can get help on the usage
of these commands (mktree, gendata and display), by typing the
command, with no arguments, at the UNIX prompt.  First, though, we
recommend that you look at the text file "sample_session", which
contains a session with OC1 illustrating some of the available 
options. Detailed descriptions of individual modules can be found
in the files mktree.readme, gendata.readme, and display.readme.

The complete, sufficiently (?!) documented, C source code of OC1 is
available with this package.

This directory also contains sample files (linear.train, sample.dt)
giving the formats of a data file and a decision tree file,
respectively.

Finally, a note: using multivariate tests at each node of a decision
tree has both advantages and disadvantages. The resulting trees may be
smaller and/or more accurate, but they may be more time-consuming to
induce than univariate trees. Afterall, nothing comes for free!

Enjoy !

-Sreerama K. Murthy
 Steven Salzberg
 Simon Kasif

 Department of Computer Science
 Johns Hopkins University
 Baltimore, MD 21218
 U.S.A.