Best score in leaderboard is 0.001799 better than sample_submission score
Data Ā·29 Apr 2021, 20:32Ā·2
How is this possible? The fact of trying to predict a value between ~-1 and ~1 and getting worse results training a model than submitting all 0s seems like this problem has no proper solution.
I am also trying to understand why my models are not performing as expected. We will try our best. By the time the competition approaches the end, we will have figured out a way to approach it.
I am also trying to understand why my models are not performing as expected. We will try our best. By the time the competition approaches the end, we will have figured out a way to approach it.
Yeah thats the spirit