Project_03
Classified Information
The Data Clustering Problem


Project 3 looks at the problem of trying to classify data. We're assuming that we many sets of data, each represented as a string of numbers. Our goal is to somehow arrange this data into groups, and perhaps discover how to use a smaller selection of the data that can represent the range of the entire set.

Reference:

  1. Nargess Memarsadeghi, Dianne O'Leary,
    Classified Information: The Data Clustering Problem,
    Computing in Science and Engineering,
    Volume 5, Number 5, September/October 2003.
  2. Dianne O'Leary,
    Scientific Computing with Case Studies,
    SIAM, 2008,
    ISBN13: 978-0-898716-66-5,
    LC: QA401.O44.


The data is stored in a single image called "Charlie". To view the image stored in the MATLAB file, issue the commands

        load charlie;
        image ( asdata / 255 );
        axis image;
        axis off;
      
For better viewing, use the command
        set ( gcf, 'Renderer', 'opengl' )
      
To work with the image in MATLAB, you will need to extract the data as an array of double precision real numbers:
        load charlie;
        asdata = double ( as );
      

You can go up one level to the Computational Science Projects page.


Last revised on 10 February 2009.