Forecasting time series with sieve bootstrap.Reportar como inadecuado




Forecasting time series with sieve bootstrap. - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Publisher: Elsevier

Issued date: 2002-01-01

Citation: Journal of Statistical Planning and Inference, 1 Jan. 2002, 1001, 1-11.

ISSN: 0378-3758

DOI: 10.1016-S0378-37580100092-1

Sponsor: We would like to thank Mike Wiper, two referees and the coordinating editor for carefully reading that greatly improved the paper. This research was partially supported by the Dirección General de Educación Superior project DGES PB96-0111 and Cátedra de Calidad BBVA.

Publisher version: http:-dx.doi.org-10.1016-S0378-37580100092-1

Keywords: Sieve bootstrap , Prediction intervals , Time series , Linear processes

Abstract:In this paper we propose bootstrap methods for constructing nonparametric prediction intervals for a general class of linear processes. Our approach uses the AR∞-sieve bootstrap procedure based on residual resampling from an autoregressive approximation to tIn this paper we propose bootstrap methods for constructing nonparametric prediction intervals for a general class of linear processes. Our approach uses the AR∞-sieve bootstrap procedure based on residual resampling from an autoregressive approximation to the given process. We present a Monte Carlo study comparing the finite sample properties of the sieve bootstrap with those of alternative methods. Finally, we illustrate the performance of the proposed method with a real data example.+-





Autor: Alonso, Andrés M.; Peña, Daniel; Romo Urroz, Juan

Fuente: http://e-archivo.uc3m.es


Introducción



Universidad Carlos III de Madrid Repositorio institucional e-Archivo http:--e-archivo.uc3m.es Departamento de Estadística DES - Artículos de Revistas 2002-01-01 Forecasting time series with sieve bootstrap. Alonso, Andrés M. Elsevier Journal of Statistical Planning and Inference, (1 Jan.
2002), 100(1), 1-11. http:--hdl.handle.net-10016-14838 Descargado de e-Archivo, repositorio institucional de la Universidad Carlos III de Madrid Forecasting time series with sieve bootstrap Andr$es M.
Alonso∗ , Daniel Peña, Juan Romo Departmento de Estad stica y Econometr a, Universidad Carlos III de Madrid, 28903 Gestafe, C=Madrid, Spain Received 21 February 2000; received in revised form 10 January 2001; accepted 30 January 2001 Abstract In this paper we propose bootstrap methods for constructing nonparametric prediction intervals for a general class of linear processes.
Our approach uses the AR(∞)-sieve bootstrap procedure based on residual resampling from an autoregressive approximation to the given process.
We present a Monte Carlo study comparing the 2nite sample properties of the sieve bootstrap with those of alternative methods.
Finally, we illustrate the performance of the proposed method with a real data example. MSC: 62M10; 62M20; 62G09 Keywords: Sieve bootstrap; Prediction intervals; Time series; Linear processes 1.
Introduction When studying a time series, one of the main goals is the estimation of forecast intervals based on an observed sample path of the process.
The traditional approach of 2nding prediction intervals for a linear time series assumes that the distribution of the error process is known.
Thus, these prediction intervals could be adversely a=ected by departures from the true underlying distribution.
For example, using a Monte Carlo study, Thombs and Schucany (1990) have shown that the standard (Gaussian) Box Jenkins method performs poorly given a skewed bimodal error distribution. Some bootstrap approaches have been proposed as a dist...





Documentos relacionados