Ανάλυση, μοντελοποίηση και πρόγνωση του δείκτη ναυλαγοράς Baltic Dry Index (BDI) με χρήση machine learning αλγορίθμων και νευρωνικών δικτύων

Κουτσούκος, Δημήτριος

Analysis, modeling and forecasting of the Baltic Dry Index (BDI) using machine learning algorithms and neural networks

Master Thesis

Author

Κουτσούκος, Δημήτριος

Date

2022-03

Abstract

In the framework of this Thesis, a literature review is carried out among articles that use machine learning algorithms and neural networks in order to predict the Baltic Dry Index (BDI). These surveys will be categorized according to their modeling work and prediction analysis. There will also be extensive reference to the way in which the bulk cargo transport system operates and to the factors that affect the Baltic dry index and consequently the freight rates, in order to understand the correlations that exist in the system and therefore determining the structure of the model that will be created to predict the index. In addition, the operation of the machine learning algorithms used in the literature to predict the index will be analyzed in order to find the algorithms that are best adapted to the specifics of the Baltic Dry index. The way in which the data will be collected, pre-processed and modeled in order to use them in the appropriate machine learning algorithms will be analyzed. Univariate / Multivariate analysis will then be performed to examine whether either approach produces a more accurate model for predicting the BDI freight rate index and finally, the most effective models for each of the analysis methods will be compared while an attempt will be made to improve the index forecasting model by testing the combination of the models that achieve optimal forecasting.

Postgraduate Studies Programme

Πληροφοριακά Συστήματα και Υπηρεσίες

Department

Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα Ψηφιακών Συστημάτων

Number of pages

150

Language

Greek

URI

https://dione.lib.unipi.gr/xmlui/handle/unipi/14273
http://dx.doi.org/10.26267/unipi_dione/1696

Collections

Τμήμα Ψηφιακών Συστημάτων

Show full item record

Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα