Locating Hidden Groups in Communication Networks Using Hidden Markov Models
- Cite this paper as:
- Magdon-Ismail M., Goldberg M., Wallace W., Siebecker D. (2003) Locating Hidden Groups in Communication Networks Using Hidden Markov Models. In: Chen H., Miranda R., Zeng D.D., Demchak C., Schroeder J., Madhusudan T. (eds) Intelligence and Security Informatics. ISI 2003. Lecture Notes in Computer Science, vol 2665. Springer, Berlin, Heidelberg
A communication network is a collection of social groups that communicate via an underlying communication medium (for example newsgroups over the Internet). In such a network, a hidden group may try to camoflauge its communications amongst the typical communications of the network. We study the task of detecting such hidden groups given only the history of the communications for the entire communication network. We develop a probabilistic approach using a Hidden Markov model of the communication network. Our approach does not require the use of any semantic information regarding the communications. We present the general probabilistic model, and show the results of applying this framework to a simplified society. For 50 time steps of communication data, we can obtain greater than 90% accuracy in detecting both whether or not their is a hidden group, and who the hidden group members are.
Unable to display preview. Download preview PDF.