[Python] What is the difference between reg:logistic and binary:logistic in XGBoost

I'm working on a regression problem using XGBoost and I want output between 0 to 1. My labels are continuous values between 0 to 1, I am getting values outside of 0 and 1 using normal regression methods. Can someone explain the key differences between these two objectives? Specifically, What is the cost function for both of them?

Continue reading...