Mobile Behaviometrics: Models and applications

The penetration of mobile devices equipped with various embedded sensors also make it possible to capture the physical and virtual context of the user and surrounding environment. Further, the modeling of human behaviors based on those data becomes very important due to the increasing popularity of context-aware computing and people-centric applications, which utilize users' behavior pattern to improve the existing services or enable new services. In this paper, we propose a new framework, Mobile Behaviometrics, to measure and quantify unique human behavioral patterns and natural rhythm that every user has when interacting with their mobile devices. After empirically study the similarity between human behavior and natural language, we introduce a language approach using well established natural language process (NLP) techniques: we convert the raw sensory data into behavior text representation as sequences of behavior labels. Each behavior label is considered as a word in the language. We then train n-gram language model on those traces based on which we are able to perform classification, prediction and anomaly detection. We apply these Behaviometric models and algorithms to several practical use scenarios. From the experimental results, we conclude that we are able to efficiently model and identify users via Behaviometrics analysis of the sensory data using the proposed techniques.