Hi,

A short question regarding the documentation for the GradientBoostingRegressor. In the documenation for the score method to this regressor we can read that it gives the R^2 value, which is defined as 1 - u/v. Further on we read that u = ((y_true - y_pred) ** 2).sum() and is called the regression sum and that v = ((y_true - y_true.mean()) ** 2).sum() and is called the residual sum. I found this a bit strange and looked it up at Wikipedia which indeed said that u must be the residual sum and v the regression sum. I though you wanted to know this minor error.

Cheers