team: n binary decision trees voting.
- the same n for each team.
- 2. minor for less false positives.
- 3. minor for voting.
- many, but balanced(?) team of trees.
- as many as possible. (Until ther are no more atttributes).
we construct each tree with id3 using the columns (attributes)
that do not appear in the previously constructed trees
- eliminate all the columns ...
- eliminate some columns ...
- in each recusive invocation of the id3 algoritm,
eliminate the experiment-attribute selected.
team proposal: (m of n)
- In a team, the tree results are weighted.
- The results are equally considered.
This work is derived from
Luis Josť Cearra Zabala
Wed Nov 7 11:06:59 CET 2001