Identifying 802.11 Traffic From Passive Measurements Using Iterative Bayesian Inference
In this paper, the authors propose a classification scheme that differentiates Ethernet and WLAN TCP flows based on measurements collected passively at the edge of a network. This scheme computes two quantities, the fraction of wireless TCP flows and the degree of belief that a TCP flow traverses a WLAN inside the network, using an iterative Bayesian inference algorithm that they developed. They prove that this iterative Bayesian inference algorithm converges to the unique Maximum Likelihood Estimate (MLE) of these two quantities. Furthermore, it has the advantage that it can handle any general -classification problem given the marginal distributions of these classes.