A 'distant' reading of A Muvra: exploring the autonomist press in the Corsican language through topic modelling

Abstract

The aim of this article is to explore the autonomist press in the Corsican language through topic modelling. In order to do so, the effectiveness of two topic modelling methods, LDA and LSA, will be compared on a trilingual corpus - Italian, Croatian and French - not subject to lemmatization, not only in order to assess the results obtained, but also to understand which of these methods comes closest to our previous knowledge on the subject, which is the result of a close reading.

Keywords

• corsica
• autonomismo
• lingua còrsa
• topic modelling
• nlp
• corsican language
• autonomism