XGBoosting Home | About | Contact | Examples

Suggested Ranges for Tuning XGBoost Hyperparameters

When tuning XGBoost hyperparameters, it’s important to search over specific ranges to find the optimal values for your dataset and problem. While the ideal values will depend on the specific characteristics of your data, the following ranges provide a good starting point for tuning the most important XGBoost hyperparameters:

    'max_depth': [3, 5, 7, 9],
    'min_child_weight': [1, 3, 5, 7],
    'subsample': [0.6, 0.7, 0.8, 0.9, 1.0],
    'colsample_bytree': [0.6, 0.7, 0.8, 0.9, 1.0],
    'learning_rate': [0.01, 0.05, 0.1, 0.2]

These ranges are based on common practices and experience in the data science community, but they are not exhaustive. Once you’ve identified the most promising region of the hyperparameter space, you can use a smaller range with more granular values for fine-tuning.

Why These Ranges are Suggested

The rationale behind each suggested range is as follows:

Remember that these ranges are based on experience and common practices, but the optimal values will depend on the specific characteristics of your dataset and problem. It’s important to experiment and adapt the ranges as needed.

Additional Considerations

When tuning XGBoost hyperparameters, consider the following:

By searching over these suggested ranges and considering these additional points, you’ll be well on your way to finding the optimal XGBoost hyperparameters for your specific dataset and problem.

See Also