Machine Learning for Forecasting: A Comparative Analysis

see the original item page
in the repository's web site and access all digital files if the item*



Machine Learning for Forecasting: A Comparative Analysis (EN)

Παπαδάκης, Κωνσταντίνος (EL)
Papadakis, Konstantinos (EN)

ntua (EL)
Κόλλιας, Στέφανος (EL)
Βουλόδημος, Αθανάσιος (EL)
Στάμου, Γεώργιος (EL)
Stamou, Georgios (EN)
Voulodimos, Athanasios (EN)
Kollias, Stephanos (EN)

masterThesis

2024-03-14
2024-07-08T09:57:06Z


This thesis investigates the performance of advanced machine learning models for time series forecasting. Prophet, N-BEATS, DeepAR, DeepVAR, and the Temporal Fusion Transformer are applied to the Electricity Load Diagrams and PEMS-SF datasets. Results are rigorously evaluated using appropriate forecasting metrics. The study highlights the strengths and weaknesses of each model in handling real-world data complexities, offering insights for choosing optimal forecasting methods based on data characteristics and problem domain. (EN)


Μηχανική Μάθηση (EL)
Πρόβλεψη Χρονοσειρών (EL)
Επιστήμη Δεδομένων (EL)
Βαθιά Μάθηση (EL)
Timeseries Forecasting (EN)
Data Science (EN)
Comparative Analysis (EN)
Machine Learning (EN)
Deep Learning (EN)

English

AILS (EL)
Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών (EL)

Αναφορά Δημιουργού-Μη Εμπορική Χρήση 3.0 Ελλάδα
http://creativecommons.org/licenses/by-nc/3.0/gr/




*Institutions are responsible for keeping their URLs functional (digital file, item page in repository site)