Una aplicación de la estimación no paramétrica al modelo lineal general con varianza no homógenea
ISSN: 0041-0241
Year of publication: 1985
Volume: 36
Issue: 2
Pages: 70-87
Type: Article
More publications in: Trabajos de estadística e investigación operativa
Abstract
In this paper we introduce a new estimate of the regression line in a heteroscedastic linear model. If the variance function is smooth we can use the nonparametric estimation for estimate it and then define a new weighted least square estimate. We show that this estimate is asymptotically optimal with minimum variance.