Hi, I am using XGBoostClassifier for Binary Classification. From the page: https://xgboost.readthedocs.io/en/latest/parameter.html, regarding the “objective” parameter, there are several options. I am confusing of their meaning:
- reg:logistic: logistic regression
- binary:logistic: logistic regression for binary classification, output probability
From the above description, I understand that if I set objective = “reg:logistic” or objective = “binary:logistic”; the model turns out to be logistic regression (kind of simple and linear model; not Gradient boosting any more). Is that true? If not, please let me know why the name “logistic regression” is used for describing reg:logistic and binary:logistic.
After all, what are possible options for XGBoost Classifier for binary classification? I want output to be score (kinds of probability, not 0/1 prediction).
Besides, what are the difference between XGBoost from https://xgboost.readthedocs.io/en/latest/parameter.html and sklearn.
I am looking forward to hearing from you.