Skip to main content
. 2021 Mar 25;11:6877. doi: 10.1038/s41598-021-86320-0

Table 4.

Hyperparameters of the XGBoost classifier.

Hyperparameter Values Hyperparameter Values
Number of trees 10, 20, 30 Feature subsampling rate 0.7 or 0.5
Tree depth 5 ASD sample weight 2
Learning rate 0.01 Lasso and ridge coefficients 10

The feature subsampling rate is set to 0.7 and 0.5 for semi-automatic and automatic feature selection strategies respectively.