Authorship analysis in cybercrime investigation

摘要: Criminals have been using the Internet to distribute a wide range of illegal materials globally in an anonymous manner, making criminal identity tracing difficult cybercrime investigation process. In this study we propose adopt authorship analysis framework automatically trace identities cyber criminals through messages they post on Internet. Under framework, three types message features, including style markers, structural and content-specific are extracted inductive learning algorithms used build feature-based models identify messages. To evaluate effectiveness conducted experimental data sets English Chinese email online newsgroup We experimented with all features algorithms. The results indicate that proposed approach can discover real authors both relatively high accuracies.