This class is returned from the loess1D function and holds the state of a loess regression with one predictor variable.
Compute the weights for robust loess, for all robustness levels <= the robustness parameter. This computation is embarrassingly parallel, so if a TaskPool is provided it will be parallelized.
Predict the value of y when x == point, using robustness iterations of the biweight procedure outlined in the reference to make the estimates more robust.
Obtain smoothed predictions of y at the values of x provided on creation of this object, for the given level of robustness. Evaluating these is computationally expensive and may be parallelized by providing a TaskPool object.