Application of Stylometry to DarkWeb Forum User Identification

The fast growth of the cyberspace in recent years has served as a convenient channel for criminals to do their illegal businesses, especially in Dark Web the hidden side of the Internet. The anonymous nature of Dark Web forums makes them ideal environments for criminal discussions. Ranging from government, security agencies to financial institutions, many parties are willing to trace the identities of the suspects through these online conversations. Dark Web participants usually have multiple accounts on various forums. On multiple occasions, being able to validate that multiple accounts on different Dark Web forums belong to the same person with high enough confidence allows us to combine various scattering pieces of information into a more concrete and advanced form of knowledge. Such knowledge will lead to actionable insights which are very useful for bringing the criminals to justice. In this paper, we examine the effectiveness of writing style analysis (stylometry) for linking multiple accounts in different Dark Web forums. Initial evaluations have shown that the proposed methodology is promisingly practicable, having a high potential to assist the investigators in exposing anonymous identities in cyber environments.