Robust nonparametric regression: Review and practical considerations

Nonparametric regression models offer a way to understand and quantify relationships between variables without having identify an appropriate family of possible functions. Although many estimation methods for these have been proposed in the literature, most them can be highly sensitive presence small proportion atypical observations training set. A review outlier robust nonparametric is provided, paying particular attention practical considerations. Since outliers also influence negatively estimator by affecting selection bandwidths or smoothing parameters, discussion alternatives this task included. Using “classical” estimators (and their counterparts) very challenging settings with moderate large number explanatory variables, so recently that scale well growing covariates are discussed.