On Semantic Organization and Fusion of Trajectory Data

With the proliferation of positioning mobile devices, people’s trajectory data are posted on the net including spatial locations and semantic contexts such as in the form of text like twitter posted text. How to organize or fuse the raw spatial trajectories and context semantic data into a structured whole for analysis further is a problem, the focus of which is mostly how to annotate episodes in raw trajectories. In this paper we examine a structured and partially self-describing way for semantic organization and fusion of trajectory data. We annotate episodes with structured sentiments, events, or topic words, where sentiments given in a self-describing way and events are represented using the form from the natural language processing literature. Besides, all the data in the whole model are represented with JSON.