Joint Bayesian Optimization Results
==================================================

Timestamp: 20260817_112204
Best optimization objective (mean_skill_huber): 0.2981
Best CV Score (mean R): 0.7159
Best mean skill ratio: 0.2981

Model Configuration:
  L0_a: XGBOOST, target_transform=standard, loss=none
  L0_b: XGBOOST, target_transform=standard, loss=none
  L1_a: XGBOOST, target_transform=standard, loss=none

XGBoost device: cpu
Inner CV stratification: GMM on training-set groups only
Sample weights: fit_weight_source=weighted [weighted_fit_weight], use_pd_weight=True
Parsimony: mode=off, strength=0.0

Pruning Parameters:
  Redundancy %: 0
  Error %: 0
  Error recursions: 0
  Samples removed: 0
  Samples kept: 2382

L0_a Hyperparameters (XGBOOST):
  n_estimators: 495
  learning_rate: 0.10384360026951815
  max_depth: 4
  min_child_weight: 10
  colsample_bytree: 0.9743321606047528
  subsample: 0.6200464904857294
  reg_alpha: 1.6885509548365776e-06
  reg_lambda: 5.058989941177468

L0_b Hyperparameters (XGBOOST):
  n_estimators: 199
  learning_rate: 0.013277613654969232
  max_depth: 5
  min_child_weight: 10
  colsample_bytree: 0.9827189637255969
  subsample: 0.8006738094567716
  reg_alpha: 0.17830292623082372
  reg_lambda: 1.9996404569961638e-06

L1_a Hyperparameters (XGBOOST):
  n_estimators: 484
  learning_rate: 0.03893207684747986
  max_depth: 9
  min_child_weight: 3
  colsample_bytree: 0.5250557635029864
  subsample: 0.8827551637701143
  reg_alpha: 2.79747650285095e-08
  reg_lambda: 1.8984753200182491

