In the analysis of real time series, especially those with high sampling frequency, it is possible that the data is influenced by more than one seasonal component. The mSARIMA models are an extension of the SARIMA models and allow us to capture the presence of multi-seasonality. The work consists of defining the class of mSARIMA models, verifying their correct functioning through a simulation study, comparing their forecasting accuracy with the most well-known models in the literature and applying it to the study of the tidal phenomenon in Venice.
Nell’analisi delle serie storiche reali, specialmente quelle con elevata frequenza di campionamento, é possibile che i dati siano influenzati da più di una componente stagionale. I modelli mSARIMA sono un’estensione dei modelli SARIMA e consentono di cogliere la presenza di multistagionalità. Il lavoro consiste nel definire la classe di modelli mSARIMA, verificarne il corretto funzionamento tramite uno studio di simulazione, confrontarne l’accuratezza previsionale con i più conosciuti modelli in letteratura e applicarla allo studio del fenomeno delle maree a Venezia.
Modelli mSARIMA per serie storiche multistagionali: analisi e applicazioni
BORTOLATO, ALBERTO
2023/2024
Abstract
In the analysis of real time series, especially those with high sampling frequency, it is possible that the data is influenced by more than one seasonal component. The mSARIMA models are an extension of the SARIMA models and allow us to capture the presence of multi-seasonality. The work consists of defining the class of mSARIMA models, verifying their correct functioning through a simulation study, comparing their forecasting accuracy with the most well-known models in the literature and applying it to the study of the tidal phenomenon in Venice.File | Dimensione | Formato | |
---|---|---|---|
Bortolato_Alberto.pdf
accesso aperto
Dimensione
7.09 MB
Formato
Adobe PDF
|
7.09 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/68527