Im building a XGBoost regression model to predict the values in the range of -3 to 3. Im using Root Mean Squared Error to evaluate the model. With hyper-parameter tuning and everything the best scores I could get was this:
- training - 0.118
- validation - 0.3207 (5 folds)
- testing - 0.3018
I know for the regression task, the RMSE values should be as less as possible. But however, I couldn't reach the range of 0.2 atleast for the test score.
Im wondering does my validation and test score range of 0.3 has anything to do with my -3 to 3 continuous target range?
Any quick help here is much appreciated. Thank you