I built a model using XGBoost algorithm to predict precipitations. It turns out that the RMSE is equal to 7.6. Does it mean that the model performs poorly? If so, what would be your piece of advice to improve it?
I actually tried to turn the model with some hyper parameters like n_estimators, learning rate, max_depth, etc. but it doesn’t decrease the RMSE value. So, I will be glad to see your comments and support.
The RMSE value is in the same units as the response variable, so it depends on the predicted answers.
For example, if your answers are close to thousands or millions, an error of 7.6 is meaningless. If your answers are around 10, an error of 7.6 is a problem.