XGBoost
Extreme Gradient Boosting — the original modern GBDT library, slightly slower but very robust.
Definition
XGBoost (Extreme Gradient Boosting), introduced by Tianqi Chen in 2014, was the first widely-adopted boosting framework with built-in regularisation, parallel training and out-of-core computation. It dominated Kaggle competitions for years and remains a strong baseline in quant finance. Compared to LightGBM, XGBoost grows trees level-wise (more conservative) and tends to generalise slightly better on small datasets. Most production ensembles blend both libraries.
Formula
Regularised boosting objective:
Obj = sum_i Loss(y_i, y_hat_i) + sum_k Omega(f_k)
Omega(f) = gamma*T + 0.5*lambda*||w||^2
T = number of leaves
w = leaf weights
gamma, lambda = regularisation strengthWorked example
XGBoost with 800 trees, max_depth 6, learning_rate 0.03, lambda 1.0 predicts month-ahead stock direction with AUC 0.58 on a held-out walk-forward fold. Combined with LightGBM via simple averaging the ensemble AUC rises to 0.61.
How ARIA Analyst uses it
ARIA's ML ensemble averages LightGBM and XGBoost predictions across 9 regime×horizon bundles, with XGBoost contributing stability and LightGBM contributing speed of retraining.
Related terms
LightGBM
A fast gradient-boosted decision tree framework from Microsoft, dominant on tabular financial data.
Isotonic Calibration
A non-parametric monotonic transformation that maps raw model scores to well-calibrated probabilities.
Purged K-Fold Cross-Validation
A cross-validation scheme that removes overlapping training samples to prevent look-ahead leakage.
Walk-Forward Analysis
A backtesting procedure that retrains the model on a rolling window and tests on the next out-of-sample period.
See XGBoost in action on any asset
ARIA Analyst computes XGBoost automatically as part of a hybrid multi-agent investment report — 5 deterministic scoring agents plus AI augmentation (ML ensemble, Bull vs Bear debate, 10 Deep Search agents on Premium). Get yours in seconds.
Try ARIA Analyst free →