Gradient Boosting

Gradient Boosting

Gradient Boosting is a very powerful learning algorithm. It works as an ensemble of different simple learning algorithms, like decision trees. The trees are not working in parallel like in a Random Forest (“bagging approach”), but sequentially, in a way that each tree tries to recognize the errors of the previous tree correctly (“boosting approach”).

Good explanations of Gradient Boosting can be found in the following articles:

Gradient Boosting from Scratch (Medium Article)

A Kaggle Master explains Gradient Boosting