Abstract—In Bowers, et al., a technique was presented, referred to as Iterative Language Translation (ILT), for reducing the threat of deanonymization attacks via two well-known author identification systems (AISs). In this paper, we introduce four additional ‘stronger’ AISs, which outperform the AISs evaluated in Bowers, et al. Our results show that ILT still remains an effective technique for reducing author identification accuracy even if stronger AISs are used.
Index Terms—Author identification, feature extraction, feature selection, steady state genetic algorithm (SSGA).
The authors are with the North Carolina A&T State University, USA (e-mail: namack@aggies.ncat.edu, jdbowers@aggies.ncat.edu, hcwillia@aggies.ncat.edu, gvdozier@ncat.edu, jashelt1@aggies.ncat.edu).
Cite: Nathan Mack, Jasmine Bowers, Henry Williams, Gerry Dozier, and Joseph Shelton, "The Best Way to a Strong Defense is a Strong Offense: Mitigating Deanonymization Attacks via Iterative Language Translation," International Journal of Machine Learning and Computing vol.5, no. 5, pp. 409-413, 2015.