Groupanizer: a method to correlate multi-users position with daily moments

Abstract

Groupanizer is an extension of groupware and constitutes a development platform to integrate user-centric information for the benefit of groupware applications. User-centric context is used in a collaborative manner to positively link and reinforce group member idiosyncrasies. In order to do so, monitoring of the user's daily context is essential, and thus, one aspect we aim at integrating is user physical position. We believe that common visited places and trajectories constitute one user-centric element that should be used to bias transactions - namely tasks or scheduling - in groupware. Although physical location could by itself be integrated in a model, we believe that motion behaviors are intrinsically linked to moments of daily life. We define those moments using three context elements - a time period, an activity and the current day. This definition leverages position mapping to include its implicit nature. Essentially, the aim is for our model to ideally reflect a group's ecology, namely relations between multiple users given their current and future locations. This paper defines the ontology, mapping locations and moments among users, and the hidden Markov model used to derive location patterns. Finally, it describes the experimentation phase and its results