I have a question concerning your Random Forest algorithm implementation. I tried to understand the code but unfortunately I wasn't able to fully uncover everything.
I suppose you have changed the algorithm in comparison to classic Random Forest algorithm proposed by Breiman (as we are talking about ranking). I just wonder if you modified the algorithm according to one of the papers like Liangxiao JIAN: Learning random forests for ranking or Stephan Clemencon: Ranking Forests? Or which changes have you applied in order to fit the algorithm to the ranking purposes?
And second question, does your implementation of Random Forests support multi-graded relevance or just relevant/irrelevant grades?
Thank you for any clarification
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
I have a question concerning your Random Forest algorithm implementation. I tried to understand the code but unfortunately I wasn't able to fully uncover everything.
I suppose you have changed the algorithm in comparison to classic Random Forest algorithm proposed by Breiman (as we are talking about ranking). I just wonder if you modified the algorithm according to one of the papers like Liangxiao JIAN: Learning random forests for ranking or Stephan Clemencon: Ranking Forests? Or which changes have you applied in order to fit the algorithm to the ranking purposes?
And second question, does your implementation of Random Forests support multi-graded relevance or just relevant/irrelevant grades?
Thank you for any clarification
Marek,
My implementation is a straight implementation of Random Forests for regression (point-wise learning to rank).
Because it does regression, it certainly supports multi-graded relevance.
Thank you! It helped me a lot!