Included in this directory are 8 files:

  0readme (this)

Data file:
  ExampleData.txt               - expression data file from Jeff Trent's Lab

Output files:
  ga_knn_info.txt               - information file
  selection_count.txt           - number of times a variable is being selected
                                  Scatter plot may be used to visulize the data.
  variable_ranked_by_GA_KNN.txt - top 500 variables in Eisen's clustering program 
                                  format. If you are interested in only the top50,
                                  take the first 51 rows (50+1 sample name row) 
                                  in the list and plug them in the Eisen's Cluster 
                                  and TreeView.
  loocv_update.txt              - result of prediction of the current left-out
                                  sample. Updated every 500 near-optiaml solutions.
  prediction_loocv.txt          - result of leave-one-out prediction using top-
                                  ranked variables from 1 to 200.

If you run the software with the following arguments:

./ga_knn -a 3 -c 1 -d 20 -f ExampleData.txt -k 3 -n 22 -p 0 -r 18 -s 5000 -t 20 -v 3226 -N 1

you should get similar results. In this example, 21 out of the 22 samples were 
correctly classified. Sample 17 [class: 3] was classified as 1. It took
approximately 1.5 hrs on a 2GHz linux machine.

A few different files will be generated if you are running multiple splits
of the data set.
