Weka Intro
----------
by Yihua Wu

Weka is a collection of machine learning algorithms for solving
real-world data mining problems.  The algorithms can either be applied
directly to a dataset or called from your own Java code.  Weka
contains tools for data pre-processing, classification, regression,
clustering, association rules, and visualization.  It is also
well-suited for developing new machine learning schemes.  Weka is open
soure software issued under the GNU General Public License.

1. Installation

Weka is installed on the CS grad machines in /grad/s3/weka323.  If you
use csh you'll need to do the following before using it:

	setenv WEKAHOME /grad/s3/weka323 .

If you use ksh or bash, do "

	export WEKAHOME=/grad/s3/weka323 .

You then need to include "$WEKAHOME/weka.jar" in the java classpath.
It is also available on the research machines under
/farm/unsupported---to use it change "/grad/s3" to
"/farm/unsupported/lib" in the preceding instructions.  You can also
download the software yourself to use on your own computer from the
webpage at http://www.cs.waikato.ac.nz/~ml/weka/index.html .

2. How to use Weka

Under directory of "weka323", there are several help files.  "README",
"README_Experiment_Gui" and "Tutorial.pdf" are where you can get al
detailed information on how to run weka.  Here are some highlights:

2.1 Data File

The machine-learning software cannot be run without data.  Under
"weka323", there is a sub-directory named "data", in which all data
files are saved as "*.arff".  When you run weka, you can either use
software included data files or define your own.  As for data format,
please refer to the second part of the README.

2.2 Running Command

Double click the weka.jar icon, or from a command line type:

	java -jar weka.jar .

(Please read README for details.)

 * Select "Preprocess" to open a data file;
 * Select "Classify" to choose a classifier (Decision Tree, Neural
   Network, etc.);
 * Press "Start", and check the output in window of "Classifier
   output".