Skip to main content
. 2021 Feb 16;12:619857. doi: 10.3389/fgene.2021.619857

TABLE 1.

Hyperparameter values used in the Random Forest and XGBoost analyses.

RF (threefold CV) XGBoost (threefold CV)
Adrenal gland Mtry = 0.1 × M Colsample = 0.01
Ntree = 2,000,000 Eta = 0.5
Hypothalamus Mtry = sqrt (M) Colsample = 0.01
Ntree = 2,000,000 Eta = 0.5
Liver Mtry = sqrt (M) Colsample = 0.01
Ntree = 1,000,000 Eta = 0.01
Muscle Mtry = sqrt (M) Colsample = 0.01
Ntree = 1,000,000 Eta = 0.01
Pituitary Mtry = 0.1 × M Colsample = 0.03
Ntree = 2,000,000 Eta = 0.01

M, total number of genes in individual tissue; Ntree, number of trees; Mtry, number of genes for forming a decision tree; colsample, proportion of genes subsampled for a decision tree; eta, learning rate; CV, cross validation.