Spatiotemporal Behavior Profiling: A Treasure Hunt Case Study

Trajectories have been providing us with a wealth of derived information such as traffic conditions and road network updates. This work focuses on deriving user profiles through spatiotemporal analysis of trajectory data to provide insight into the quality of information provided by users. The presented behavior profiling method assesses user participation characteristics in a treasure-hunt type event. Consisting of an analysis and a profiling phase, analysis involves a timeline and a stay-point analysis, as well as a semantic trajectory inspection relating actual and expected paths. The analysis results are then grouped around profiles that can be used to estimate the user performance in the activity. The proposed profiling method is evaluated by means of a student orientation treasure-hunt activity at the University of Twente, The Netherlands. The profiling method is used to predict the students' gaming behavior by means of a simple team type classification, and a feature-based answer type classification