##################################################################
     Gene Selection and Sample Classification Using
  the Genetic algorithm/k-nearest neighbors algorithm

                  Author: Leping Li

   National Institute of Environmental Health Sciences
           Research Triangle Park, NC 27709
              Email: Li3@niehs.nih.gov

               Copyright (c) 1999-2003

   ---------------------Version 1.02---------------------  
        Last modification: December 18, 2003

Date & time calculation performed: Tue Sep 21 19:09:46 2004


Data info:
   Total number of samples:                   22
   Number of samples in training set:         21
   Number of samples in test set:             1
   Number of variables (genes or m/z...):     3226
   Data file:                                 ExampleData.txt

GA parameters:
   Number of niches:                          3
   Number of generations:                     50
   Population size:                           50
   Chromosome length (d):                     20
   Termination fitness cutoff:                18
   Number of solutions specified:             5000

Others:
  Random seed number:                         1095808186

Total number of classes:    3, individual class type:   1   2   3 .
Number of samples in each class:   7 [class - 1]   8 [class - 2]   7 [class - 3] 
Minimal and maximal numbers of samples in a class:    7    8

  class[1]:    1    2    3    4    5    6   18 
  class[2]:    7    8    9   10   19   20   21   22 
  class[3]:   11   12   13   14   15   16   17 

KNN:
  k-nearest neighbors= 3
  A majority rule applies.
    - A majority of the neighbors must agree that is AT LEAST
      2 out of the 3 neighbors must be the same type.

Application 3...
   A leave-one-out cross-validation (LOOCV) is carried out.
  
Information on data processing:
    Data are log2 transformed.

##################################################################
