User profiling and re-identification: Case of university-wide network analysis
| Authors | |
|---|---|
| Year of publication | 2009 |
| Type | Article in Proceedings |
| Conference | Trust, Privacy and Security in Digital Business, 6th International Conference, TrustBus 2009 |
| MU Faculty or unit | |
| Citation | |
| Doi | http://dx.doi.org/10.1007/978-3-642-03748-1_1 |
| Field | Informatics |
| Keywords | user profiling; network analysis; data mining; IDF; similarity searching; cosine similarity |
| Description | In this paper we present our methodology for context information processing, modeling users' behaviour and re-identification. Our primary interest is to what extent a user can be re-identified if we have his "user profile" and how much information is required for a successful re-identification. We operate with ``user profiles'' that reflect user's behaviour in the past. We describe the input date we use for building behavioural characteristics; similarity searching procedure and an evaluation of the re-identification process. We discuss (and provide results of our experiments) how different initial conditions, as well as different approaches used in the similarity searching phase, influence the results and propose the optimal scenario where we obtain the most accurate results. We provide experimental results of re-identification of three protocols (SSH, HTTP and HTTPS). |
| Related projects: |