XGBoost
In plain terms
A fast, high-performance gradient-boosting library.
Going deeper
XGBoost implements optimized gradient boosting with regularization and speed tricks, dominating tabular ML competitions. It's efficient and accurate.
Under the hood
Parallelized tree building, regularization, and handling of missing values make it a go-to for structured data.
Examples
Kaggle wins, tabular prediction, fraud detection, ranking, structured data.
Think of it like...
Like a finely-tuned racing version of the gradient-boosting engine.