Πρόβλεψη ταχύτητας πλοίου με χρήση χρονοσειρών
View/
Keywords
Spatiotemporal data ; Time series ; Vessel speed prediction ; VAR model ; ARIMA modelsAbstract
This research has been conducted during my postgraduate degree at the University of Piraeus, Department of Digital Systems and the main goal is to predict vessel’s speed with timeseries. More specifically, the data that have been used was downloaded from the following link: Heterogeneous Integrated Dataset for Maritime Intelligence, Surveillance, and Reconnaissance | Zenodo, and contain data that were collected from the West coast of France, the Bay of Biscay, with AIS sensors, for time period of six months, from October 1st, 2015 to March 31st, 2016. ARIMA and VAR models have been used for our predictions. For the timeseries analysis we created one dataset for ARIMA model that contains vessel speed by datetime and one dataset for VAR model that contains ship, weather, wind and ocean data by datetime. The criteria that have been selected to evaluate the predictions of models are AIC, BIC, RMSE and MAE. After analyzing and comparing these two models, we conclude that the VAR model performs better than ARIMA with RMSE score of 2.67 knots.
Postgraduate Studies Programme
Πληροφοριακά Συστήματα και ΥπηρεσίεςDepartment
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα Ψηφιακών ΣυστημάτωνNumber of pages
64Language
GreekCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα