Air Iuav archivio istituzionale della ricerca
AIR è l'Archivio Istituzionale della Ricerca ad accesso aperto dell'Università Iuav di Venezia che ha lo scopo di raccogliere, documentare, diffonderee conservare i prodotti scientifici dell'ateneo.
AIR IUAV si avvale di IRIS, il sistema di gestione integrata dei dati della ricerca del CINECA.
EnglishMotivated by the need of extracting local trends and low frequency components in non-stationary time series, this paper discusses methods of robust non-parametric smoothing. Basic approach is the combination of the parametric M-estimation with kernel and local polynomial regression methods. The result is an iterative estimator that retains a linear structure, but has kernel weights also in the direction of the prediction errors. The design of smoothing coefficients is carried out with robust cross-validation criteria and rules of thumb. The method works well both to remove the influence of patches of outliers and to detect the local breaks and persistent structural change in time series.
| Titolo: | Robust non-parametric smoothing of non-stationary time series |
| Autori: | |
| Data di pubblicazione: | 2009 |
| Rivista: | |
| Abstract: | Motivated by the need of extracting local trends and low frequency components in non-stationary time series, this paper discusses methods of robust non-parametric smoothing. Basic approach is the combination of the parametric M-estimation with kernel and local polynomial regression methods. The result is an iterative estimator that retains a linear structure, but has kernel weights also in the direction of the prediction errors. The design of smoothing coefficients is carried out with robust cross-validation criteria and rules of thumb. The method works well both to remove the influence of patches of outliers and to detect the local breaks and persistent structural change in time series. |
| Handle: | http://hdl.handle.net/11578/31883 |
| Appare nelle tipologie: | 1.1 Articolo su Rivista |
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Utilizza questo identificativo per citare o creare un link a questo documento:
http://hdl.handle.net/11578/31883
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
Copyright © 2020