Exploring Predictability in Mobile Interaction

Abstract

Recent endeavors in mobile computing concentration analyzing the predictability of human behavior by means of mobility models synthesized from real mobile user traces. Currently, the main focus of such studies is physic allocation: discovering travel patterns, estimating real user movements and anticipating the whereabouts and dynamics of individuals. In this paper, we propose to widen the analyzed context as to take into account a more natural activity inhuman behavior, namely interaction. As such, we explore the predictability of user synergy based on tracing data collected from mobile phone users in academic and office environments. We take into account interactions over Bluetooth and over wireless networks and, by measuring the entropy of interacting both with peers and wireless access points, we discover a remarkable invariability in synergic patterns.