Συσταδοποίηση δεδομένων κίνησης με χρήση βαθιών νευρωνικών δικτύων

View/ Open
Keywords
Συσταδοποίηση ; Δεδομένα κίνησης ; Νευρωνικά δίκτυα ; Βαθιά νευρωνικά δίκτυα ; Ομαδοποίηση ; Βαθιά γνώσηAbstract
The rapid development of wireless infrastructure and data acquisition technologies has contributed to the exponential increase in the volume of data to be processed specifically for trajectory data which records the positions of moving objects by time or period and is now an important means of study of human behavior or solving traffic problems such as traffic problems. In this study we focus on a recent work that uses deep learning methods to embed the original space of motion data and then grouping it with the K-means algorithm. After we implemented it, we developed it in the grouping part through 4 grouping algorithms with different characteristics. Our results showed that the selection of K-means in the original work was not optimal.