Authorship attribution and text clustering in contemporary Italian novels

The aim of authorship attribution methods is to exploit observable textual features to identify the distinctive traits of an author’s writing style. In quantitative approaches, the problem of authorship attribution is often tackled in terms of the search for a reliable procedure to measure the similarity between texts and for a consistent text clustering method. Among several proposals available in the scientific literature, this new study explores a large corpus of contemporary Italian novels and focusses on a famous case of disputed authorship. The object of the research is Elena Ferrante's work, with a specific focus on the presumed similarity with other novels written by Italian authors from Naples and its surroundings. Although we expected that the authors' geographical origin should play a role in terms of topic selection and linguistic features, results show that regional traits are not particularly relevant: of all the 10 candidates taken into consideration, the only author who can be considered “textually close” to Elena Ferrante is Domenico Starnone.