Monotone constraints in vector-leaf models

[jprafael]

Both monotone constraints and vector-leaf models can act as regularizers by limiting the allowed splits.

• Monotone constraints allow embedding domain knowledge into the model and prevent noise-derived splits.
• Vector Leaf models prefer splits that are good for many targets also preventing noise-derived splits.

However using both at the same time is sometimes impossible with the current API.

Consider the case where a feature has positive effect for target 1 and negative effect for target 2. This is a very common scenario, for example with multi-class classifiers. Because we can only supply a single monotone_constraints vector we can't describe these contradicting effects to the model. It is however trivial when training multiple trees (pass different monotone_constraints for each train).

Is there any planned work on making these two features work in tandem?

[trivialfis]

I don't have a clear idea on how to do it yet. The monotone constraint is specialized for the vector leaf model. (See the monotone constraint tutorial) Would you like to create an issue to help keep track of the feature request?

[jprafael]

I wasn't sure if creating an issue for a feature request would be appropriate. Doing it now

[RAMitchell]

I think we need vector constraints.

[jprafael]

Closing in favor of issue 12506