I have optimised the model as per the lecture, and I came up with:
Test RMSE: 0.48
Test R^2: 0.56
OOB score: 0.87
Train RMSE: 0.09
Train R^2: 0.98
This is against the Valid.csv in Kaggle
Unfortunately I don’t know how to test against the Test.csv, because in Kaggle the solution is not published, so I dont have the y_test for that.
Regardless, would low R2 but high (log) RMSE and OOB score make the cut?
My notebook: https://bit.ly/2ML2FgH