I am using XGBoost to predict a variable that is highly skewed and always is greater than zero. I did a significant search to see some materials for gamma objective function in XGBoost but I could not find anything being understandable for me. Considering that I have two options: 1. Use that gamma objective function without completely understanding it. 2. Transform the target variable to gamma space and fit a model.
Do you think that the second method is logical and it helps me to improve the performance of the model? I also wonder if you share some information about how gamma's objective function works.