any update about the XGBoost and LightGBM implementations?
I am new to RM. I have been looking for XGBoost in RM but apparently it is not implemented. I find it very strange that it is not implemented since it is a very popular and useful algorithm, or is there some other way to implement it in RM?
Thanks,
Aaron ST
Find more posts tagged with
Sort by:
1 - 2 of
21
MartinLiebig
Altair EmployeeAccepted Answer
Hi @yoreme ,
we have it on our radar and if I am not mistaken XGBoost is the example for our python operator framework. So you can simply create your own python based operator to run XGBoost. You can check the help of Python Learner for some more infos: https://docs.rapidminer.com/latest/studio/operators/modeling/python_learner.html
So far its a big hurdle to integrate LightGBM and XGBoost, since they are C(++) based. We are a Java platform, so this is not that straight forward. We carefully monitor this of course. If there is enough traction for alternative, 'native' XGBoost/Catboost/lightGBM implementations we will of course check what we can do.
Best,
Martin
New Altair Community Member
Please not that the first tutorial process for the Python Learner operator already wraps LightGBM. If you have Python and LightGBM installed, it should work out of the box (you might have to configure your Python environment in RapidMiner first).
Sort by:
1 - 1 of
11
Altair EmployeeAccepted Answer
Hi @yoreme ,
we have it on our radar and if I am not mistaken XGBoost is the example for our python operator framework. So you can simply create your own python based operator to run XGBoost. You can check the help of Python Learner for some more infos: https://docs.rapidminer.com/latest/studio/operators/modeling/python_learner.html
So far its a big hurdle to integrate LightGBM and XGBoost, since they are C(++) based. We are a Java platform, so this is not that straight forward. We carefully monitor this of course. If there is enough traction for alternative, 'native' XGBoost/Catboost/lightGBM implementations we will of course check what we can do.
Best,
Martin