1. CONICET
  2. Producciónes CyT
Producción CyT

THE USE OF THE DURBIN-WATSON TEST IN NON-LINEAR REGRESSION MODELS

Articulo

Autoría:

ABRIL, JUAN CARLOS ; Maekawa, Koichi

Fecha:

1999

Editorial y Lugar de Edición:

ISOSS PUBL

Revista:

PAKISTAN JOURNAL OF STATISTICS, vol. 15 (pp. 41-46) ISOSS PUBL

Resumen *

In this paper we show that the original Durbin-Watson test can be used in non-linear cases following a very simple procedure. We show as well that this will be valid to the second order of approximation. Firstly we present the case of a single independent variable and a single parameter of the non-linear model. Then we treat the p dimensional case. Finally we give some concluding remarks. Información suministrada por el agente en SIGEVA

Palabras Clave

Asymptotic expansionsSerial correlationNon-linear regressionDurbin-Watson test statistics