Fan fictions are an important part of today's reading culture and have gained more and more interest as data material in the Natural Language Processing community. However, fan fictions themselves have become objects of research in various research areas in the humanities like gender studies, fan studies, literary studies and internet studies. Furthermore, various projects in Digital Humanities explore fan fictions via computational text analysis. One interesting point is the question on how the fan community transforms the original work when creating fan fictions.
In this thesis, we want to investigate the transformation process between original and fan created fan fictions via methods of computational text analysis on large-scale corpora. As case study for the analysis, we select the popular tv show „Supernatural“ which is among the most popular and famous material for fan fictions. To investigate differences between original and fan fictions we look at the following variables: * character mentions and character networks * sentiment and emotion expression * frequencies of various word types like gender specific words * other metrics of intertextuality
Experience in