Robust Nonlinear Regression in R
Hossein Riazoshams, Habshah Midi, Gebrenegus Ghilagaber · 2018
There are many nlr packages in the R language comprehensive archive for robust nonlinear regression. For comparison of the packages, this chapter shows a simulation study, because the exact values are known and the biases can therefore be computed. The nlrq function from the nlrq package fits a nonlinear regression model by quantile regression. The chapter then illustrates one easy and one complicated examples: lakes data example, and simulated data example. The nlrob function in the robustbase package fits a nonlinear regression by iteratively reweighted least squares. The chapter also shows the quantile regression, least median squares (LMS), and ordinary least squares (OLS) estimates. The estimates from nlrq and nlrob are close to the OLS estimate computed by the nlr and nls functions.