PROJECT TITLE :

Modeling Spatial Trajectories with Attribute Representation Learning

ABSTRACT:

The widespread use of positioning devices has resulted in the generation of a large number of trajectories, each of which possesses four attributes: user ID, location ID, and time-stamp, as well as an implicit attribute known as activity type (which is analogous to the concept of "topic" in text mining). Existing works learn different attribute representations in order to model these trajectories. This can be done by either introducing latent activity types based on topic models or by transforming the location and time context into a low-dimensional space using embedding techniques. Both of these options are available. In this paper, we propose a comprehensive method that we will refer to as the Human Mobility Representation Model (HMRM). This method will simultaneously produce the vector representations of all four characteristics, both explicit and implicit. The benefits of using HMRM include the following: (1) it models the latent activity types and learns trajectory attribute embeddings in an integrated manner; and (2) it connects the activity-related distributions and these attributes embeddings by adding a newly designed collaborative learning component, and makes them mutually exchanged in order to take advantage of the best of both worlds. On two real check-in datasets obtained from Foursquare, we apply HMRM to both unsupervised and supervised tasks, including two activity evaluation tasks and two embedding evaluation tasks. These tasks are comprised of two activity evaluation tasks and two embedding evaluation tasks. According to the findings of the experiments, HMRM has the potential to not only improve performance in terms of capturing latent activity types but also learn better trajectory embeddings.