cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Choose Language Hide Translation Bar

Hyperparameter search in Predictive Modeling platforms

In the XGBoost add-in (https://community.jmp.com/t5/JMP-Add-Ins/XGBoost-Add-In-for-JMP-Pro/ta-p/319383), there is a very nice option where using DoE, a hyperparameter search can be launched, to identify the best combination of hyperparameters to maximize a given objective function.

 

dmmdiego_0-1642106316678.png

 

This option is extremely helpful, and it's a standard approach when training models in R or Python.

I think there would be tremendous value implementing the same option for other predictive modeling platforms, such as Neural, Boosted Tree, Boostrap Forest and Support Vector Machine. In each case, an experimental matrix could be setup to define the search ranges for the selected hyperparameters.

The XGBoost add-in accomplishes this very nicely, so you guys already figured out the mechanics, it's a good time to make it available for other models!

1 Comment
kayleewarren
Staff
Status changed to: Acknowledged

@dmmdiego Thank you for your suggestion! We have captured your request and will take it under consideration.