VizStruct Handbook

 
You are hereTutorial (2)Tutorial (2)Tutorial (2)Tutorial (2)
Environment System: SUN Solaris 9
Web Server:  Apache HTTP server 2.0.50
CGI: Perl 5.8.4
Programming: MATLAB R14 and MATLAB Webserver 2.0
File Format Numeric data matrix without missing values.
  1. Plain text format. Each entry can be either decimal of scientific expression. No missing entry is allowed. No string is allowed
  2. If the original data has missing values. They should be filled using '999'. VizStruct can use algorithm to fill these values
  3. The leading column (first column) may contain data class labels. This is optional. By default, the leading column is data value. The class labels must be integers starting with '1'. However, these lables are not necessarily consecutive integers
Demo Files
  1. iris-unsort.txt: original data set. size 150 x 4, 3 classes (1) setosa 1-50, (2) versicolor 51-100, (3) virginica 101-150
  2. iris.txt: iris data set with columns have been sorted using alternating cosine algorithm
  3. msifn.txt: array data from 14 multiple sclerosis patients prior to and 24 hours after IFN treatment. size 28 x 88
  4. golub.txt: array data from 72 leukemia patients of two subtypes: acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML). size 72 x 50. This is the combination of training (27 ALL and 11 AML) and testing (20 ALL and 14 AML) data set
  5. alon.txt: array data from 22 normal and 40 colon cancer patients. size 62 x 50
  6. stuart.txt: measurements of gene expressions during rat kidney organogenesis, data set consists of 873 genes which vary significantly during kidney development at 7 different time points. 5 groups of genes have been distinguished. They are aggregated together. The number of these genes are 323,70,47,262 and 166
  7. stuart-label.txt: stuart.txt with the leading column contains class labels numbered from 1 to 5
  8. cho.txt: data set about mRNA transcript levels during the cell cycle of the budding yeast S. cerevisiae. It consists of 416 genes containing 17 time points. Five groups of genes are reported by the author: 67,135,75,54 and 88
Contact VizStruct is maintained by Li Zhang (lizhang@cse.buffalo.edu)
Tutorial Tutorial(1)Lets begin -> Tutorial



© 2001 Li Zhang. All right reserved. Department of Computer Science and Engineering
State University of New York at Buffalo
Last Modified: 9/6/2004