Abstract (eng)
The purpose of this master’s thesis is to apply natural language processing (NLP) methods to a corpus of 182 Gothic fiction texts in order to gain insight into the genre’s composition and explore its early influences in a network. To this end various computational approaches had been employed including machine learning models, exploratory data analysis and clustering, yet focusing on network analysis, and topic modeling. The results are aggregated and compared across different categories as well as throughout time. The consistent stock of characters and motifs within Gothic Fiction, lends itself well for an analysis of its recurrent composing elements. This quality had made it a frequent topic of explorations within traditional schools of literary criticism, such as Structuralism and Formalism. This thesis aims to expand on previous approaches with the tool set of the Digital Humanities and Distant Reading, taking a more quantitative perspective, while arguing for the necessity of closer future collaboration between digital and analogue research in the humanities, in order to enable new avenues of investigation.