My question is whether the returned training loss values in R package xgboost version 1.2.0.1 contain regularization terms. Specifically, in the following example, does train-logloss contains regularized term related to gamma, lambda and alpha? If not, how do I obtain objective value with regularized term? Could someone point to relevant code in the R package xgboost? Thanks!
x=matrix(rnorm(100*2),100,2)
g2=sample(c(0,1),100,replace=TRUE)
library("xgboost")
fit2=xgboost(x, g2, nrounds=5, objective="binary:logistic", eval_metric="logloss", gamma=1, lambda=1, alpha=1)
[1] train-logloss:0.646722
[2] train-logloss:0.591888
[3] train-logloss:0.530950
[4] train-logloss:0.520571
[5] train-logloss:0.491039