How do you tune XGBoost at scale without overfitting?hard

Type
conceptual
Topic
do-you-tune-xgboost-at-scale-without-overfitting
Frequency
common
Tags
machine-learning, how, you, tune, xgboost, scale
Answer

Key params: max_depth (keep 3-6), min_child_weight (higher = less overfit), subsample and colsample_bytree (0.7-0.9), small learning_rate + more rounds, lambda/alpha for L2/L1 reg.

Explanation

Key params: max_depth (keep 3-6), min_child_weight (higher = less overfit), subsample and colsample_bytree (0.7-0.9), small learning_rate + more rounds, lambda/alpha for L2/L1 reg. Use early stopping on a validation set. For scale: Bayesian optimization (Optuna) instead of grid search. Always monitor train vs val loss gap.

Follow-upCan you give a production example?