What is the default loss criterion that CARET's random forest uses for classification (e.g. gini, entropy, log-loss)?
In scikit-learn, gini
is the default loss function of the random forest (source). However, I cannot find what the default loss function is for CARET's random forest. I checked the CARET source code and documentation, but could not find a clear answer there. Any suggestions would be really helpful.
Thanks!