ETD

• Agglomerative
• Clustering
• Data
• Distances
• Hierarchical
• Mobility
• Points
• Similarity
• Trajectory

Understanding trajectory data is instrumental in extracting a pattern from moving objects, this can be applied in several areas such as urban planning, intelligent transportation, and so on. In this thesis, we consider the spatial information for our clustering algorithm. Our contribution to the trajectory clustering includes creating distance functions such as discrete Frechete, DTW, EDR, ERP, Hausdorff, and a distance matrix function to perform our matrix computation. In addition, we use the hierarchical agglomerative clustering method to group our precomputed matrix into clusters and validate our clustering by using the silhouette score or silhouette coefficient. The final clusters from the trajectory clustering are visualized in a map representation. In our clustering we consider the different periods of the day (morning, afternoon, etc.) and when the taxis are occupied or not.