Robust Estimators for Partially Linear Additive Models

Robust estimation for partially linear additive models based on B-splines. The package implements a robust proposal that combines robust linear MM-regression estimators, which use an initial S-estimator followed by an M-estimation step with fixed scale, and B-spline approximations. The M-estimation step uses Tukey's bisquare function. The least-squares counterpart is also implemented. See Boente and Martínez (2023) for more details. The package also provides robust and least-squares estimators for additive models, that is, when the covariates enter only the nonparametric component of the model.