Extreme Gradient Boosting Algorithm, abbreviated as “XGBoost”, is a specialized form of Decision Trees (KA) and stands out in the literature as a classification, prediction and ranking method.

The XGBoost algorithm, which was used by all the top 10 solutions selected in the 2015 Knowledge Discovery and Data Mining (KDD) cup, attracts the attention of researchers from all branches and is a very effective and complex popular method in terms of the results it provides. The XGBoost algorithm, which avoids overfitting with the regulation factor and adopts a multi-threaded approach that learns supervised and leads to more speed and performance by using the machine's CPU core appropriately, promises potential for all sectors and constitutes an important field of study for researchers with its open-to-development structure.

In this study, the evolution of the XGBoost algorithm to date, its application experience in the literature (studies conducted - literature review), its content, structure, operation, parameters, its difference from the Gradient Boosting Algorithm (GAA), its advantages and disadvantages are examined and it is aimed to provide a general idea about the method to interested researchers.