Andrushko, TarasTarasAndrushko2024-05-082024-05-082024https://fis.uni-bamberg.de/handle/uniba/93378This poster discusses the technical aspects of compiling a corpus of TikTok comments documenting (dynamics in) language attitudes towards Ukrainian, Russian and Surzhyk, especially problems in dealing with trilingual data. On preparing the data for our study of language attitude, several computational linguistics problems had to be solved, namely: working with multilingual texts and recognizing sarcasm (which can also be multilingual).engUkrainischNatural Language ProcessingTikTokKorpus490articleurn:nbn:de:bvb:473-irb-933785