Answers are grounded in this paper and cite their sources.
Consider the regression model Yᵢ = g(tᵢ) + εᵢ, 1 ≤ i ≤ n, with nonrandom design variables (tᵢ) and measurements (Yᵢ) for the unknown regression function g(·). We assume that the data are heteroscedastic, i.e., E(ε²ᵢ) = σ²ᵢ ≡ const. and investigate how to estimate σ²ᵢ. If σ²ᵢ = σ²(tᵢ) with a smooth function σ²(·), initial estimators σ̃²ᵢ can be improved by kernel smoothers and the resulting class of estimators is shown to be uniformly consistent. These estimates can be used to improve the estimation of the regression function g itself in parametric and nonparametric models. Further applications are suggested.
No takes yet. Share an insight, caveat, or question.
Müller et al. (1987) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: