Uncovering an Author's Regional Background through High-Resolution Authorship Profiling Using Social Media Data

In an international online world it is easy to disguise one’s identity. Methods in forensic linguistics and stylometry work well to uncover authors from a set of candidates, but break down when no candidates are available for comparison. However, writing carries information that can be used to gather intelligence about an author: We can employ methods of authorship profiling, the assessment of linguistic features to infer author characteristics like age or gender. This knowledge can be used, for example, to filter a list of suspects or assess the veracity of authors’ claims about their identity. Profiling the regional background of authors has received limited attention in the literature, especially in a forensic context. Thus, in this talk I present the state of the art in high-resolution regional profiling in the German-speaking area. The data in this study consists of 21 million social media posts from the platform Jodel. First, I will demonstrate that the corpus can be used to map and identify regional patterns using lexis. This information can then be leveraged to help identify the regional backgrounds of authors. In light of these results I will also consider a more general application of the proposed method and tools.