Gradient Boosting Regressor component - ClickML Studio

Gradient Boosting Regressor builds trees iteratively, where each tree corrects the residual errors of the current ensemble. It typically achieves the best accuracy among regression models in the toolbox.

Configuration

Parameter	Description	Default
N Estimators	Number of boosting stages.	`100`
Learning Rate	Contribution of each tree. Lower = more conservative, more trees needed.	`0.1`
Max Depth	Maximum depth of individual trees.	`3`
Subsample	Fraction of samples used per tree. Below 1.0 adds stochasticity.	`1.0`
Random State	Seed for reproducibility.	`42`

Input / Output

	Type
Input	`X Train` + `Y Train`
Output	Trained Model

When to use

Best when accuracy is the priority and training time is acceptable. Does not require feature scaling.