Moving Taylor Bayesian Regression (MoTaBaR) is a method for nonparametric multi-dimensional function estimation with correlated errors from finite samples, developed at the Potsdam Institute for Climate Impact Research.
