Hi!
I am new to Machine Learning and was trying my hand at different models and ensembles when I faced a small issue.
Having used the scikit-learn package before trying xgboost, I got used to using the model.score() method to get a the mean of the accuracy scores, R2, etc.
On using it in the xgbRegressor model, I still got the scores. I wasn’t sure what the default scoring metric was so I opened the docs. It was then I realized that the score method is not documented in the XGBRegressor Documentation.
It has also not been implemented in the XGBoostRegressor. And I found out that XGBRegressor uses the XGBRegressorBase class which in turn uses RegressorMixin by scikit-learn which has the score() method implemented. So I am really not sure if this was intended or a mistake.
Please let me know your thoughts on this. I would like to work on this issue
TLDR;
the score method is not documented but still works due to the RegressorMixin class.