Here are some notes taken from the following seminar

TITLE:
     Application of data mining techniques on survey data

SPEAKER:
     Ms Supunmali Ahangama (supunmali@comp.nus.edu.sg)
     Doctoral Candidate
     Department of Information Systems, School of Computing,
     National University of Singapore

DATE/TIME:
November 29, 2013, 10:30am – 12noon, Fri

VENUE:
     Executive Classroom (COM2-04-02)
     School of Computing, National University of Singapore

Abstract:
Data mining isnt exclusively the purview of expensive software. There are
software tools like R and Weka, capable of performing all the same things as
expensive software.  According to 2011 Data miner survey carried out by Rexer
Analytic, R was growing in its popularity and by 2011 it was used by close to
half of all the data miners (47%) participated in the survey. Weka is another
free tool with a graphical user interface and it provides an API to use Weka
libraries in your Java applications. This presentation aims to provide a basic
introduction to R and Weka and the applicability of data mining techniques in
studying consumer behavior.

Biodata:
Ms Supunmali Ahangama is currently a Doctoral Candidate in the Department of
Information Systems, School of Computing at the National University of
Singapore. Her research interests include health analytics, eHealth and mobile
computing.

R is one the most popular tool for Data mining. Currently No. 2 most popular tool. (No. 1 is STATISTICA)

To use the data mining interface, you can install Rattle by typing the following three commands.

  • install.packages(“rattle”)
  • library(“rattle”)
  • rattle()

After installation, you see the following screen

image

 

First, one needs to import the data into the program, then you can use the “Explore” tab to explore the data. You can also use Neural Network in the Model tab to use artificial neural network to explore the data in R.

 

You can also explore the same data in Weka using explore tab.

image

 

Once the data is imported, you may need to convert the data from numerical to normal. Then, can go to classify tab to use MultiLayerPerception (Neural Network), and choose classify (cross-validation) to perform sensitivity analysis.

The talk did not reveal too much details. For more information about how to use Weka, you may want to refer to this tutorial:

https://blog.itu.dk/SPVC-E2010/files/2010/11/wekatutorial.pdf

Written by Shengdong Zhao

Shen is an Associate Professor in the Computer Science Department, National University of Singapore (NUS). He is the founding director of the NUS-HCI Lab, specializing in research and innovation in the area of human computer interaction.