Algorithm 3.

Weight optimization fitness function.

input the individual x

begin

1. partition the training data T into 4 equal

partitions Tj, j = 1 . . . 4

2. S = 0

3. forj = 1 → 4 do

begin

4.  train the base learners with T − Tj

5.  test the base learners with Tj

6.  use the weights in x to aggregate the base learners

and store the number of correctly classifified cases in s

7.  S = S + s

end

8. returnS

end