Αποτίμηση ομοιότητας ολιστικών τροχιών κινούμενων αντικειμένων
Evaluation of similarity in holistic trajectories
![Thumbnail](/xmlui/bitstream/handle/unipi/13721/Georgakas_18006.pdf.jpg?sequence=4&isAllowed=y)
View/ Open
Keywords
Moving objects ; Semantic trajectories ; Trajectory transformations ; Similarity ; Holistic trajectories ; Κινούμενα αντικείμεναAbstract
For many years the research on evaluation of similarity of trajectories has focused on two-dimensional sequences of points, considering only space and time information, called raw trajectories. However, the increase use of GPS technologies and social media, has led many approaches to enrich these moving objects with multiple dimensions including semantic dimension. As a result, most recent approaches have proposed similarity measures that support space, time and semantics. A way to compare the effectiveness and the robustness of these similarity measures is to transform the trajectories and compute the similarity between the seed trajectory and the transformed trajectories. In our thesis, we propose a method that applies different types of transformations over semantic trajectories and creates sets of moving objects, according to a predefined rate r. These transformations, have the ability to transform the information of space, time and as well as all the attributes of the semantic dimension. For each set, we calculate the similarity of the original trajectory in relation to the transformed trajectories, using different similarity measures.