Define the fitted value of ! , from a regression with n observations. If you are given a value of the regressor as %'# (that did not appear in any observation in your original sample), what would be your prediction for the outcome %'# ? Present your answer algebraically and graphically