An Introduction to Weka with Demos

Tomasz Oliwa

Ph.D., Computer Science, The University of Georgia, USA

Dipl.-Inform., University of Koblenz-Landau, Germany..

Lecture in CSCI/ARTI 8950 Machine LearningSlides at: http://www.cs.uga.edu/~tomasz/weka2013fall/

September 5, 2013

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 1 / 18

Weka Software

The Weka machine learning/data mining suite:

Rich collection of preprocessing, (un)/supervisedlearning and evaluation algorithms

Accessible GUI

Java, Free software (GPL):http://www.cs.waikato.ac.nz/~ml/weka/

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 2 / 18

Attribute-Relation File Format (ARFF)

@relation weather

@attribute outlook sunny, overcast, rainy

@attribute temperature real

@attribute humidity real

@attribute windy TRUE, FALSE

@attribute play yes, no

@data

sunny,85,85,FALSE,no

sunny,80,90,TRUE,no

overcast,83,86,FALSE,yes

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 3 / 18

Weka Modes1 Explorer: Direct application and visualizations2 Experimenter: Run & compare algorithms3 Knowledge Flow: Visual composition of workflow4 Simple CLI: Command line interface5 Source code: Weka classes in Java code

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 4 / 18

Weka GUI Chooser

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 5 / 18

Preprocessing

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 6 / 18

Preprocessing - Load Data

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 7 / 18

Preprocessing - Data Overview

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 8 / 18

Preprocessing - Visualize All

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 9 / 18

Classification

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 10 / 18

Classification Results

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 11 / 18

Classification Results - DT

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 12 / 18

Clustering Results

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 13 / 18

Attribute Selection Results

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 14 / 18

Demos

Demo time:

Preprocessing

Classification

Regression

Clustering

Feature Selection

Experimenter

Source code import

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 15 / 18

Take Home Message

Weka is a powerful machine learning suite:

Explorer: Intuitive suite with result visualizations

Experimenter: Batch perform statistical tests onresults

Source code: Directly import Weka classes in yourprograms

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 16 / 18

Questions

Thanks for your attention!

Do you have any questions?

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 17 / 18

References and Tutorials

Weka: http://www.cs.waikato.ac.nz/~ml/weka/

Weka Wiki: http://weka.wikispaces.com/

IBM Weka Tutorials:

Introduction and Regression:http://www.ibm.com/developerworks/opensource/

library/os-weka1/index.html

Classification and clustering:http://www.ibm.com/developerworks/opensource/

library/os-weka2/index.html

Nearest Neighbor and server-side library:http://www.ibm.com/developerworks/opensource/

library/os-weka3/index.html

Tomasz Oliwa (oliwa@uga.edu) Introduction to Weka and Demos September 5, 2013 18 / 18