Σχεδιασμός συστήματος συλλογής δεδομένων, κατασκευή προγραμματιστικών διεπαφών και προβλεπτικών μεθόδων για δεδομένα καταστροφών

Βαγιανος, Ευάγγελος; Vagianos, Efangelos

dc.contributor.author	Βαγιανος, Ευάγγελος	el
dc.contributor.author	Vagianos, Efangelos	en
dc.date.accessioned	2023-01-23T11:47:58Z
dc.date.available	2023-01-23T11:47:58Z
dc.identifier.uri	https://dspace.lib.ntua.gr/xmlui/handle/123456789/56835
dc.identifier.uri	http://dx.doi.org/10.26240/heal.ntua.24533
dc.rights	Αναφορά Δημιουργού - Παρόμοια Διανομή 3.0 Ελλάδα	*
dc.rights.uri	http://creativecommons.org/licenses/by-sa/3.0/gr/	*
dc.subject	Disaster Databses	en
dc.subject	Machine learning	en
dc.subject	JSON	en
dc.subject	Python	en
dc.subject	Jupyter	en
dc.subject	Notebook	en
dc.subject	Βασεις Δεδομενων καταστροφης	el
dc.subject	Αλγοριθμοι μηχανικης μαθησης	el
dc.title	Σχεδιασμός συστήματος συλλογής δεδομένων, κατασκευή προγραμματιστικών διεπαφών και προβλεπτικών μεθόδων για δεδομένα καταστροφών	el
heal.type	bachelorThesis
heal.classification	Databases Machine learning	en
heal.language	el
heal.access	free
heal.recordProvider	ntua	el
heal.publicationDate	2022-10-25
heal.abstract	This thesis deals with the issue of searching and selecting heterogeneous disaster databases. The disasters analyzed, are divided into two minor categories according to the cause of their creation: these categories are geophysical and weather-related. Initially, databases containing a lot of data on these disasters had to be identified. Many government agencies and companies provide free access to all users to process their databases and draw useful conclusions. Nine databases were selected, and their contents and validity were compared. The criteria primarily concern the complete documentation and completeness of these databases. It was observed that many bases which apparently met the criteria were rejected due to missing records or because they contained outdated data. The nine selected bases were extensively analyzed based on their specific features. Then, an additional sorting (discarding) was done to select the five most complete. Using Python Pandas, these bases were freed from the redundant information they contained and configured based on a specific desired pattern. Using JSON, the database data were presented after the necessary processing had been done in Jupyter. The systems were analyzed in detail how they are used. Five machine learning (ML) classification algorithms were chosen and analyzed extensively by using definitions and examples. In the ‘’Disaster_Size’’ column of the Global landslide Catalog base, all five were applied to compare their effectiveness and to draw conclusions about the time and probability of success of each model.	el
heal.advisorName	Ασκούνης, Δημήτριος	el
heal.committeeMemberName	Ψαρρας, Ιωαννης	el
heal.committeeMemberName	Δουκας, Χρυσοστομος	el
heal.committeeMemberName	Ασκουνης, Δημήτριος	el
heal.academicPublisher	Εθνικό Μετσόβιο Πολυτεχνείο. Σχολή Ηλεκτρολόγων Μηχανικών και Μηχανικών Υπολογιστών. Τομέας Ηλεκτρικών Βιομηχανικών Διατάξεων και Συστημάτων Αποφάσεων	el
heal.academicPublisherID	ntua
heal.numberOfPages	84 σ.	el
heal.fullTextAvailability	false