Tom Bylander
Department of Computer Science
Weka is a collection of machine learning algorithms that is implemented in Java. Weka includes algorithms for classification, numeric prediction, clustering, and association rules. This document provides information about the installation and usage for faculty and students, especially for the Data Mining course.
Weka is installed on on the CS Sun Network in the directory /home/bylander/weka. Documentation for weka is in this directory and in the doc subdirectory. The documentation includes the following:
a README file for installation and rudimentary use,
another README file for using a GUI for Weka.
a comprehensive tutorial describing how to run the algorithms in Weka.
the objects and methods implemented in Weka.
many datasets that can be input to Weka algorithms.
This describes how to run Weka in combination with DB2.
When you do:
as suggested by the DB2 document, it does bad things to your CLASSPATH environmental variable. I suggest that you put the following sequence of commands in your .cshrc file.source /db2home/db2inst1/sqllib/db2cshrc
This should get you all the Java objects for Weka and DB2, though I have to admit that I haven't tried to use any of the DB2 Java.source /db2home/db2inst1/sqllib/db2cshrc setenv WEKAHOME /home/bylander/weka unset CLASSPATH setenv CLASSPATH .:/home/bylander/weka/weka.jar:/db2home/db2inst1/sqllib/java/sqlj.zip:/db2home/db2inst1/sqllib/function:/db2home/db2inst1/sqllib/java/db2java.zip:/db2home/db2inst1/sqllib/java/runtime.zip
For ksh, all you should need to do is to add:
to your .profile file, but I haven't tested this.WEKAHOME=/home/bylander/weka CLASSPATH=/home/bylander/weka/weka.jar:$CLASSPATH
The README file has several examples of commands to try.
Last updated on