uclamii / model_tuner

A library to tune the hyperparameters of common ML models. Supports calibration and custom pipelines.
Apache License 2.0
3 stars 0 forks source link

Boosting Variable Rename #45

Closed elemets closed 2 weeks ago

elemets commented 4 weeks ago

Rename variable xgboost_early to reflect both CatBoost and XGBoost usage

Description

The current variable xgboost_early is now used for both CatBoost and XGBoost models, but the name is specific to XGBoost. To avoid confusion and improve clarity, we need to rename this variable to something more general that reflects its functionality for both models.

Tasks

  1. Rename the variable:

    • Change the variable xgboost_early to a more general name like early_stopping or boost_early_stopping to indicate that it works for both CatBoost and XGBoost.
  2. Update code:

    • Replace all occurrences of xgboost_early in the codebase with the new variable name.
    • Ensure compatibility in both CatBoost and XGBoost model training pipelines.
  3. Update tests:

    • Refactor the existing tests to use the new variable name.
    • Ensure all relevant tests for early stopping in both CatBoost and XGBoost still pass.
  4. Update documentation:

    • Modify any references to xgboost_early in the documentation to reflect the new name.
    • Ensure the updated documentation clearly explains that the variable is applicable to both CatBoost and XGBoost.

Notes

elemets commented 2 weeks ago

There is also a variable feature_names that is not used at all and needs to be cleaned up.

elemets commented 2 weeks ago

These have both been removed and updated