Aristotle University Of Thessaloniki (AUTH)
Επιχειρησιακή σύντηξη δεδομένων για τη βελτίωση της εκτίμησης της ποιότητας του αέρα με μεθόδους μηχανικής μάθησης
Abstract
dc:descriptionAtmospheric pollution constitutes one of the most critical environmental challenges in contemporary cities with significant implications for the health of their inhabitants. Addressing this issue necessitates accurate estimation of pollutants at high spatial resolution. Existing methods for assessing air quality present limitations and challenges. The aim of this work is to improve the spatial resolution and reliability of air quality estimates in cities, under operational conditions. Utilization of new technologies, the abundance of freely available data, and Machine Learning are used in an approach towards data fusion in the direction of smart cities. A wide range of methods is tested in two urban environments with different conditions (Thessaloniki, Helsinki). Additionally, a new method (GAHS) is developed that successfully combines information from different models, with potential applications to many similar problems. Another significant finding of the work is the improvement of the accuracy of low-cost air quality sensors as a network. Results demonstrate the capability of the data fusion approach to upgrade the spatial resolution of information at the city-block level while achieving higher accuracy compared to other available methods for air quality estimations in each area. Finally, all the above are tested under operational conditions and, in the case of Thessaloniki, an information system is designed and implemented which provides real-time estimates through a web application.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Kassandros, Theodosios
- Κάσσανδρος, Θεοδόσιος
Subjects
dc:subject × 12Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/55137