Discover User Behaviour from Trajectory as Polygons (TaP)
Ting Wang · International Journal of Applied Physics and Mathematics · 2014
In this paper, we propose a new algorithm, namely Trajectory as Polygons (TaP). Using the legacy Convex Hull algorithm (6), together with a sliding time window mechanism, TaP uses polygon to represent objects trajectory, instead of using line segments like many existing works do. The mobility patterns and user behavior can be observed from the geometric properties (e.g. location, size, shape, and number of vertices/edges etc.) of these polygons. We note that TaP is not only a solution to one single specific problem, but also a general method to treat and represent locational data, to discover information and extract knowledge-regardless the quality and density of the source data. For example, we will demonstrate how peoples' lifestyle patterns can be extracted classified and clustered using TaP even when the source data is scarce and largely inaccurate. A brief introduction to the existing works and challenges are discussed in Section II, followed by the introduction of the TaP algorithm in Section III. We talk about how to use TaP to study trajectories and study user behavior in Section IV, and a case study of how TaP could be used with locational data to discover people's lifestyles is discussed in Section V. Section VI concludes the paper with the strength of TaP and future research directions.