Extreme Gradient Boosting Algorithm for Predicting Shear Strengths of Rockfill Materials

Table 2

Parameter configuration.

Algorithm	Parameter optimization
XGBoost	n estimators = 40, learning rate = 0.250, maximum depth = 4
SVM	Cost = 8, regression loss epsilon = 0.1, kernel type = radial basis function
RF	Number of trees = 15, limit depth of individual trees = 3
KNN	Number of neighbors = 5, metric = euclidean, weight = uniform
AdaBoost	Number of estimators = 2, learning rate = 0.1, boosting algorithm = SAMME, regression loss function = linear