Privacy and E-Authentication: The Dangers of Self-disclosure in Social Networks

摘要

We propose a Bayesian model of privacy in e-authentication and develop associated entropy-based metrics. A major contribution of this work is the application of weighted entropy to characterize the user's privacy preferences. Further, we model the effects of side information on privacy and relate it to self-disclosure on Internet web sites and social networks. Specifically, our empirical study of Internet users' information disclosure habits within social networks along with the theoretical results provide insights into building a regulatory framework to address privacy concerns in e-authentication.