E-Locus - Institutional Repository of the University of Crete - Εφαρμογή τεχνικών ανάλυσης μεγάλων δεδομένων σε προβλήματα μεταφορών και αστροφυσικής

Home Εφαρμογή τεχνικών ανάλυσης μεγάλων δεδομένων σε προβλήματα μεταφορών και αστροφυσικής

Results - Details

[Add to Basket]

Identifier

000410076

Title

Εφαρμογή τεχνικών ανάλυσης μεγάλων δεδομένων σε προβλήματα μεταφορών και αστροφυσικής

Alternative Title

A big data analytics application in transportation and astrophysics

Author

Αντωνίου, Νεκταρία Χ.

Thesis advisor

Ζέζας, Ανδρέας

Reviewer

Παπαδάκης, Ιωσήφ
Παυλίδου, Βασιλική

Abstract

The main subject of this thesis is the development of a complete computational framework for «Big Data» analytics. This framework is able to perform the pre-processing and manipulation of data, as well as their analysis. For this reason, the computational framework requires two stages: the pre-processing and preparation stage of the data and the stage of data analysis. The first stage utilizes Hadoop platform, while the second stage utilizes Artificial Neural Networks (ANN). Initially, a detailed description of the tools that are used in the computational framework is given. Then, the application of the presented framework is given on a transportation problem and an astrophysics problem. For the transportation problem, a large amount of data from all US airports were retrieved and data for one particular airport were extracted, in order to identify mean flight delays per day. Then, these data were used to predict the time required for airport operations to return back to their scheduled times (i.e., normalization time). For the second problem, an attempt was made, via ANN, to predict the morphology of galaxies based on spectral data from Sloan Digital Sky Survey (SDSS). The results from these two applications are presented and discussed extensively. Finally, conclusions are drawn on the ability of the computational framework to handle successfully problems related to Big Data analytics and suggestions for its further improvement are made.

Language

Greek

Subject

Airports

Artificial neural networks

Flights delays

Hadoop

Hive

Morphology of galaxies

Αεροδρόμια

Καθυστέρηση πτήσεων

Μέσα μεταφορών