In my test results I get an RMSE of 0.040 but in here the result is in the digits of 9.1. I know that the model can overfit and things like that, but im curious if the results are indeed in RMSE and what results are you getting in your test dataset?
sorry for late response, I hope that you found an answer, but you test results are leaking so bad. you may want to try cross validation instread of train-test-split.
and yes the leaderboard is calculated in RMSE, and the public testset of everyone is just normal.
sorry for late response, I hope that you found an answer, but you test results are leaking so bad. you may want to try cross validation instread of train-test-split.
and yes the leaderboard is calculated in RMSE, and the public testset of everyone is just normal.