Sure, it used to be a very popular question in IBM's interviews. So, a regression model overall tries to be significantly better than just the average value of the target column, but when it's even worse than the average value of the target column, the Rsq is negative; indicating this is not a good model at all.
May I know y the test r2 of adaboost is -ve?
Sure, it used to be a very popular question in IBM's interviews. So, a regression model overall tries to be significantly better than just the average value of the target column, but when it's even worse than the average value of the target column, the Rsq is negative; indicating this is not a good model at all.