How does Tree Ensemble Learner (Regression) deal with missing values? It does not seem to complain at all while Python’s RandomForestRegressor does not move forward until NaN, infinity, missing values are fixed.
Hi @tims -
If you look at the node description for the related Tree Ensemble Learner, there is a small bit about this:
The missing value handling corresponds to the method described here. The basic idea is that for each split to try and send the missing values in every possible direction and the one yielding the best results (i.e. largest gain) is then used.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.