• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 1
  • Tagged with
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • 1
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
1

[en] IDENTIFICATION OF BOX AND JENKINS: A COPARISON BETWEEN FACE AND PADÉ APPROXIMATION / [pt] IDENTIFICAÇÃO DOS MODELOS BOX E JENKINS: UMA COMPARAÇÃO ENTRE O MÉTODO FACE E O MÉTODO DE APROXIMAÇÃO DE PADÉ

LUIZ CLAUDIO RIBEIRO 18 September 2006 (has links)
[pt] Desde de 1970, quando Box e Jenkins introduziram os modelos ARMA para análise e previsão de séries temporais, muitos estudos foram desenvolvidos buscando encontrar um método mais eficiente de identificação de tais modelos. Tal fato se deu porque o método por Box e Jenkins, baseado na função de auto-correlação parcial (FACP) não são eficientes quando os modelos apresentam componentes auto- regressivas (AR) e médias móveis (MA). Estudos comparativos realizados anteriormente mostraram que dentre os métodos de identificação já desenvolvidos, o que se mostrou mais eficiente foi o baseado na função de auto-correlação extendida (FACE) de TIAO e TSAY (1992) Recentemente, Kuldeep Kumar introduziu na literatura um método de identificação baseado na teoria de aproximação de Padé. O objetivo deste trabalho é comparar o método da FACE com o método baseado na teoria de aproximação de Padé. / [en] Since 1970, when Box and Jenkins first introduced the ARMA models to analysis and predict of time series data, a lot of studies have been developed to find an efficient identification method for such models. This was due the fact that the identification method proposed by Box and Jenkins, based on Auto-correlation Function (ACF) and Partial Auto-correlation Function (PACF), are inefficient when the models have auto regressive - AR- and moving average - MA- components. Comparative studies undertaken, have shown that, among the identification methods already developed, the method based on the Extended Auto-correlation Fuction of Tiao and Tsay (1982) is the most efficient. More recently, however, Kuldeep Kumar has introduced in the literature an identification method based on the theory of Padé aproximation. The objective of this paper is to compare the Extended Auto-correlation Function method with the method based on the Theory of Padé approximation.

Page generated in 0.0459 seconds