PRINCÍPIOS E APLICAÇÕES DE REGRESSÃO LOCAL

• Autor(a) principal: Marquetti, Adalmir
• Data de Publicação: 2009
• Outros Autores: Viali, Lori
• Tipo de documento: Artigo
• Idioma: por
• Título da fonte: Análise Econômica (Online)
• Texto Completo: https://seer.ufrgs.br/index.php/AnaliseEconomica/article/view/10808
• topic: Análise não-paramétrica. Regressão local. Modelagem local. Ajustamento de curvas e superfícies. Loess. Lowess.; JEL Classification; C14; Nonparametric analysis. Local regression. Local modeling.; JEL Classification; C14

This paper presents the basic principles of local regression, a nonparametric procedure to estimate curves and surfaces through smoothing. In the nonparametric analysis, the relationship between the dependent and independents variables is estimated without reference to a priori functional form, allowing that the data speak for themselves. Local regression might be employed to study models with one or more independent variables as well as to estimate derivatives. Moreover, its statistical properties have been studied, permitting to realize inference over the results. Local regression has been implemented in a series of softwares, allowing that the researchers have access to this technique.