Modeling and prediction multivariate geostatistical techniques can be successfully applied to study the temporal behaviour of several correlated time series. In particular, in the time domain, by using variogram-based tools the analyst can easily a) identify trend and periodicity which characterize each time series, b) fit a properly Multivariate Linear Temporal (MLT) model to multiple correlated time series, c) predict the variable of interest (primary variable) at some time points after the last available observation, by taking into account the fitted model as well as the auxiliary information coming from the secondary variables. In this paper the convenience of performing a complete analysis of multiple correlated time series on the basis of geostatistical tools is illustrated through a case study concerning three environmental variables. As regards the computational aspects, a new version of the GSLib Cokb3d routine has been implemented for prediction purposes.
(2019). Multivariate geostatistical tools for time series modeling and prediction [poster communication - poster]. Retrieved from http://hdl.handle.net/10446/146847
Multivariate geostatistical tools for time series modeling and prediction
2019-01-01
Abstract
Modeling and prediction multivariate geostatistical techniques can be successfully applied to study the temporal behaviour of several correlated time series. In particular, in the time domain, by using variogram-based tools the analyst can easily a) identify trend and periodicity which characterize each time series, b) fit a properly Multivariate Linear Temporal (MLT) model to multiple correlated time series, c) predict the variable of interest (primary variable) at some time points after the last available observation, by taking into account the fitted model as well as the auxiliary information coming from the secondary variables. In this paper the convenience of performing a complete analysis of multiple correlated time series on the basis of geostatistical tools is illustrated through a case study concerning three environmental variables. As regards the computational aspects, a new version of the GSLib Cokb3d routine has been implemented for prediction purposes.File | Dimensione del file | Formato | |
---|---|---|---|
GRASPA2019_pp78-81.pdf
accesso aperto
Versione:
publisher's version - versione editoriale
Licenza:
Creative commons
Dimensione del file
275.93 kB
Formato
Adobe PDF
|
275.93 kB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
Aisberg ©2008 Servizi bibliotecari, Università degli studi di Bergamo | Terms of use/Condizioni di utilizzo