Article Preview
TopTrajectory Pattern Mining
Recently, the research area on trajectory data mining has grown a lot. Studies on this area consist in analyzing the mobility patterns of moving objects and in identifying groups of trajectories sharing similar patterns. In last years, many methods and techniques for trajectory pattern discovering have been proposed to meet a broad range of applications. Zheng (2015) presents a systematic survey on the major research into trajectory data mining and classifies existing patterns in four categories: (1) Moving together patterns; (2) Clustering; (3) Frequent sequence patterns; and (4) Periodic patterns. This work focuses on the first category and propose a new approach to identify moving together objects.
Examples of patterns that discover a group of objects that move together for a certain period are flock (Gudmundsson & van Kreveld, 2006; Tanaka, Vieira, & Kaster, 2015; Vieira, Bakalov, & Tsotras, 2009), group (Taniar & Goh, 2007; Yida Wang, Lim, & Hwang, 2006), convoy (Jeung, Yiu, Zhou, Jensen, & Shen, 2008), swarm (Li, Ding, Han, Kays, & Nye, 2010), traveling companion (Tang et al., 2012), gathering (Zheng, Zheng, Yuan, & Shang, 2013; K. Zheng, Zheng, Yuan, Shang, & Zhou, 2014) and co-movement (Fan, Zhang, Wu, & Tan, 2016). Moving together patterns are useful for a high number of applications, such as monitoring of delivery trucks (Jeung et al., 2008) and identification of vessels that fish together.