TESTS FOR INDEPENDENCE IN NONPARAMETRIC REGRESSION

 

  • John Einmahl – Tilburg University – The Netherlands
  • FCUL (DEIO) – Campo Grande – Bloco C/6 Piso 4 – Sala 6.4.30- 14h 30m
  • Quarta-feira, 19 de Março de 2008
 
 Abstract: Consider the nonparametric regression model Y = m(X) + E, where the function is smooth, but unknown. We construct tests for the independence of and X, based on n independent copies of (X,Y). The testing procedures are based on differences of neighboring Y´s. We establish asymptotic results for the proposed tests statistics, investigate their finite sample properties through a simulation study and present an econometric application to huosehold data. The proofs are based on delicate empirical process theory.

JEL classifications. C12, C14, C52                                                                                                        

Short running title. Tests for independence                                                                                              

 Key words and phrases. Empirical process, model diagnostics, nonparametric regression, test for independence, weak convergence.