Τεχνικές αποθήκευσης δεδομένων & εξόρυξης γνώσης για βάσεις κινούμενων αντικειμένων
Μαρκέτος, Γεράσιμος Δ.
Analyzing mobility data that are collected from location aware devices enables us to discover behavioral patterns that can be explored in applications like service accessibility, mobile marketing and traffic management. Online analytical processing (OLAP) and data mining (DM) techniques can be employed in order to convert this vast amount of raw data into useful knowledge. Their application on conventional data has been extensively studied during the last decade. The high volume of generated mobility data arises the challenge of applying analytical techniques on such data. In order to achieve this aim, we have to take into consideration the complex nature of spatiotemporal data and thus to extend appropriately the two aforementioned techniques to handle them in an efficient way. This thesis proposes a framework for Mobility Data Warehousing and Mining which consists of various components (actually, Knowledge Discovery & Delivery steps). More specifically, Trajectory Data Warehousing techniques are addressed focusing on modeling issues, ETL processes (trajectory reconstruction, data cube loading) and OLAP operations (aggregation etc). Moreover, we propose data mining techniques that explore mobility data and extract a) interaction patterns for spatiotemporal representation, synthesis and classification and b) traffic patterns that can provide useful insights regarding the traffic flow on a road network.