residuals.model_estimation

[DavidBarke]

[Tim-TU]

Since we are calculating the coefficient of determination (in case of rank_regression) one could also provide the residuals. For teaching purposes it might be useful to show that linear model assumptions aren't held.

We have to check how the R squared is calculated (i.e with respect to X or Y). We are doing a rank regression on X, so the residuals should consider this, too.

For consistency between rank_regression and ml_estimation we could provide the s3 residuals.* method for both.