search

henryzord / ardennes

An Estimation of Distribution Algorithm for Decision-Tree Induction.
5 stars 1 forks source link

One variable per level #25

Closed henryzord closed 7 years ago

henryzord commented 7 years ago

Change the GM so there is only one variable per level, as opposed to the current 2^h variables (where h is the current height of the level, starting from 0).

This should make the generated trees better (regarding their accuracy), and also decreasing the sampling procedure.

henryzord commented 7 years ago

Implemented in 74a20717e305d8dcf0782a8cbbd461e4249274e8, although it is not working for +3 levels.