Discovering and Predicting Temporal Patterns of WiFi-Interactive Social Populations

Abstract

Extensive efforts have been devoted to characterizing the rich connectivity patterns among the nodes (components) of such complex networks (systems), and in the course of development of research in this area, people have been prompted to address on a fundamental question: How does the fascinating yet complex topological features of a network affect or determine the collective behavior and performance of the networked system? While elegant attempts to address this core issue have been made, for example, from the viewpoints of synchronization, epidemics, evolutionary cooperation, and the control of complex networks, theoretically or empirically, this widely concerned key question still remains open in the newly emergent field of network science. Such fruitful advances also push the desire to understand (mobile) social networks and characterize human social populations with the interdependent collective dynamics as well as the behavioral patterns. Nowadays, a great deal of digital technologies are unobtrusively embedded into the physical world of human daily activities, which offer unparalleled opportunities to explosively digitize human physical interactions, who is contacting with whom at what time. Such powerful technologies include the Bluetooth, the active Radio Frequency Identification (RFID) technology, wireless sensors and, more close to our interest in this paper, the WiFi technology. As a snapshot of the modern society, a university is in the coverage of WiFi signals, where the WiFi system records the digital access logs of the authorized WiFi users when they access the campus wireless services. Such WiFi access records, as the indirect proxy data, work as the effective proxy of a large-scale population's social interactions.